点击显示 收起
Institute of Medicine and Section for Epidemiology and Medical Statistics, Department of Public Health and Primary Health Care, University of Bergen
Department of Thoracic Medicine and Center for Clinical Research, Haukeland University Hospital, Bergen, Norway
ABSTRACT
The effects of pre- or postnatal passive smoking on the adult incidence of asthma have not been reported previously. Between 1985 and 1996/1997, we conducted an 11-year community cohort study on the incidence of asthma and respiratory symptoms in Western Norway. The cohort included 3,786 subjects aged 15 to 70 years, of which 2,819 were responders at both baseline and follow-up. The incidence of asthma and five respiratory symptoms by self-reported exposure to maternal smoking in utero and in childhood, as well as smoking by other household members in childhood, was examined. After adjustment for sex, age, education, hay fever, personal smoking, and occupational exposure, maternal smoking was associated with asthma, phlegm cough, chronic cough, dyspnea grade 2, attacks of dyspnea, and wheezing, with odds ratios (95% confidence intervals [CI]) of 3.0 (1.6, 5.6), 1.7 (1.1, 2.6), 1.9 (1.2, 3.0), 1.9 (1.2, 3.0), 2.0 (1.3, 3.0), and 1.4 (0.9, 2.2), respectively. The adjusted attributable fractions (95% CI) of the adult incidence of asthma were 17.3% (5.2, 27.9) caused by maternal smoking and 9.3% (95% CI, eC23.2, 33.2) caused by smoking by other household members. Exposure to pre- and postnatal smoking carries a substantial risk for developing adult asthma and respiratory symptoms.
Key Words: asthma; cohort study; passive smoking; respiratory symptoms
In the last three decades, there has been a rise in asthma prevalence among children (1) and probably among adults (2), sometimes referred to as an "asthma epidemic." Some decades before the asthma epidemic, the Western world experienced a "tobacco epidemic," first among men, then among women. Although direct evidence that active smoking causes asthma is scant, several interesting findings suggest that exposure to maternal smoking in utero and/or environmental tobacco smoke (ETS) in childhood may play a part. Both pre- and postnatal passive smoking have been linked with respiratory symptoms and asthma in childhood (3eC5). Among never-smoking adults, ETS has been associated with an increased risk of respiratory symptoms and asthma, both at the workplace (6eC8) and at home (9, 10).
Finally, three recent studies have examined the effects that pre- or postnatal passive smoking in childhood may have on adult respiratory health (11eC13). In a cross-sectional study from sterbro, in Sweden, the prevalence of adult asthma was significantly higher among never-smoking subjects reporting childhood ETS (13). Similar findings were reported from cross-sectional analyses in the European Community Respiratory Health Survey, where the prevalence of asthma and wheeze were higher in adults who reported exposure to maternal smoking in utero or in childhood (11). In the one previous prospective study, data on parental smoking from the Midspan Study in Scotland, conducted between 1972 and 1976, were used to predict lung function measured in offspring in 1996 (12). In this study, maternal smoking was a risk factor for lower lung volumes (FEV1 and FVC), irrespective of the offsprings' own smoking.
In none of these three studies can one differentiate whether the outcomes occurred in childhood or adulthood. If the outcome occurred in childhood, it would confirm previous studies on pre- or postnatal passive smoking and ill effects in childhood. However, if the outcomes occurred in adulthood, it would imply that pre- or postnatal passive smoking creates a lasting susceptibility for later respiratory disease.
With the Hordaland County Cohort Study, we had a unique opportunity to examine the incidence of respiratory symptoms and asthma among adults, by passive smoking in utero or in childhood, after adjustment for several important confounders. Thus, the main objective of the current study was to examine the effects of pre- or postnatal passive smoking on the adult incidence of asthma and respiratory symptoms, both in terms of risk and preventive potential. Some of the results of this study have previously been reported in the form of abstracts presented at the European Respiratory Society conference in September 2004 (14, 15).
METHODS
Study Population
The baseline survey was conducted in 1985. A random sample of the population in the city of Bergen and 11 surrounding municipalities in Hordaland County, Western Norway, aged 15 to 70 years, received a mailed questionnaire with 40 questions about respiratory health, allergies, smoking habits, and occupational exposure (16). After two reminder letters, 3,370 subjects (89.0%) had responded. Within the 11 years of follow-up, 189 subjects died, leaving 3,181 subjects eligible for the follow-up survey, which was conducted between September 1996 and May 1997. The questionnaire was expanded to 58 questions, adding questions on education and exposure to passive smoking. A total of 2,819 subjects (88.6%) returned the questionnaire after a maximum of two reminder letters and one telephone reminder. The procedure of data sampling and collection both at baseline and follow-up has been described in detail previously (16eC18).
Questionnaire
The wording of the questions on respiratory symptoms and asthma has been described in detail previously (18). The questions have previously been validated against a British Medical Research Council questionnaire (19) and against lung function (20). Asthma was defined through the question: "Have you ever been treated by a doctor or been hospitalized for asthma"
At follow-up, the subjects also were asked three questions about their parents' smoking habits: "Did your mother smoke when she was pregnant with you," "Did your mother smoke when you were a child," and "Did others in the household smoke when you were a child" The information about passive smoking is difficult to validate in retrospect. To our knowledge, there is no biochemical method to verify lifetime exposure to passive smoking.
Statistical Analyses
The cumulative incidence of a single symptom or asthma was defined as the number of new cases in 1996/1997 divided by the number of subjects in 1985 not having the symptom or asthma (21). For the incidence of each symptom or asthma, a logistic regression model was used to estimate the adjusted odds ratios (ORs). All three passive-smoking variables were used as explanatory variables. Because of the colinearity between maternal smoking in pregnancy and childhood, only one of these two variables could be retained in each model at the same time. To further explore the difference in effect from maternal smoking in pregnancy and childhood, a combined variable was used, consisting of the following four categories: no maternal smoking, only postnatal maternal smoking, only prenatal maternal smoking, and both pre- and postnatal maternal smoking. For all logistic models, adjustment was made for the confounders of sex, age (as a continuous variable), educational level (in 1996/1997), active smoking in terms of both smoking habits and pack-years (up until 1996/1997), occupational exposure to dust or fumes (in 1985), and hay fever (in 1985).
All first-order interactions between the exposure(s) in question and the other explanatory variables were estimated through separate analyses. With the large number of potential interactions, deciding the appropriate significance level is difficult. With a Bonferroni-type correction, there is a chance of overlooking important interactions, whereas with a higher significance level, there is a chance of false-positive interactions. The interactions were tested with a chosen significance level of both 0.01 and 0.1.
The attributable fraction (AF) was defined as the proportion of the incidence in the total population related to the exposure (22). Adjusted AFs (23) with confidence intervals (CIs) were estimated according to the methodology described by Greenland and Drescher (24). Statistical analyses were conducted with Stata 8.0 (Stata Corporation, College Station, TX) (25).
RESULTS
About 1 in 10 participants reported that their mothers had smoked while pregnant with them, and a little less than a quarter reported that their mothers had smoked when they were children (Table 1). Approximately 60% reported having been exposed to ETS from other household members. For all ETS exposures, the exposure was more commonly reported in the youngest age group, among current to current smokers, and among the ex- or current smokers with the largest number of pack-years smoked (Table 1). Subjects with the lowest educational level were less likely to report exposure to passive smoking, compared with subjects with a secondary or university-level education. However, among subjects older than 40 years at baseline, exposure to passive smoking was more common among subjects with a university-level education.
The 11-year cumulative incidences of adult asthma and respiratory symptoms by passive smoking in utero or in childhood are given in Table 2. Among subjects reporting maternal smoking in pregnancy, the incidence of asthma and all symptoms were higher, compared with unexposed subjects. Among those exposed to maternal smoking in childhood, the incidence was higher for asthma and all symptoms, except chronic cough and dyspnea grade 2. The same trend was observed with exposure to smoking from other household members; however, there were smaller differences in the cumulative incidences between those exposed and those not exposed.
The results from the logistic regression analyses are shown in Table 3. Even after extensive adjustment for confounders, subjects reporting smoking exposure in utero had a significantly higher risk for the adult incidence of asthma and all symptoms, except wheezing. A weaker trend was seen with exposure to maternal smoking in childhood, with significantly increased risk for the incidence of asthma and wheezing. With the combined variable, it is apparent that the association between maternal smoking and adult incidence of asthma and respiratory symptoms is strongest with prenatal exposure. After having taken all confounders and maternal smoking into account, exposure to smoking from other household members did not show a significant relationship with the incidence of asthma and the symptoms, although the trend pointed toward a small increase in risk.
In all analyses, we adjusted for active smoking both in terms of smoking habits up until follow-up and pack-years. However, we also conducted analyses on never-smokers only, to see if the relationships held up. With a smaller sample, the CIs were widened; however, the trend was actually toward a stronger association between passive smoking in utero or in childhood and the incidence of asthma in adulthood. The OR (95% CI) for the adult incidence of asthma was 5.7 (1.9, 17.4) among never-smokers who reported maternal smoking in pregnancy.
For the incidences of the symptoms, however, there were no major differences in the estimated ORs among never-smokers only compared with the full sample with adjustment for active smoking.
When estimating all first-order interactions between the exposure in question and the confounders (sex, age, educational level, smoking habits, pack-years, occupational exposure, and hay fever), none were found to be statistically significant, with a nominal significance level of 0.01. With a significance level of 0.1, seven significant interactions were found (see online supplement). These interactions did not show a specific pattern (26). See Tables E1 and E2 in the online supplement for results.
The adjusted AFs of incident asthma and symptoms caused by ETS are given in Table 4. The AF of ETS by mother represents the fraction of incident cases caused by both pre- and postnatal tobacco exposures. Although 17.3% of incident cases of adult asthma could be explained by exposure to maternal smoking, 9.3% could be attributed to exposure to smoking by other household members. The combined total exposure to ETS in childhood could explain almost a quarter of incident cases of adult asthma. For the respiratory symptoms, the AFs varied between 3.1 and 19.0% for the total ETS exposure in childhood.
DISCUSSION
In the Hordaland County Cohort Study, subjects reporting maternal smoking in pregnancy had a marked increase in risk for developing asthma and respiratory symptoms in adulthood. Furthermore, ETS in childhood was associated with an added risk for the adult incidence of asthma and respiratory symptoms, more so if the mother smoked than if other household members smoked. The estimated adjusted AFs suggest that almost a quarter of incident cases of adult asthma could be prevented if children were not exposed to pre-and postnatal ETS.
To our knowledge, this is the first study to show that pre- and postnatal passive smoking induces a lasting vulnerability for developing asthma or respiratory symptoms. The strengths of the Hordaland County Cohort Study are its high response rate, random selection of study subjects from a general population with a wide age span, and the possibility of extensive adjustment for important confounders like personal smoking, educational level, and occupational exposure.
However, there are three important methodologic issues to consider when interpreting the results. First is the inherent colinearity between maternal smoking in pregnancy and childhood. A substantial portion of women will give up smoking when they learn they are pregnant, or while trying to become pregnant. After the pregnancy, several restart smoking. The opposite situation is rarely true; in this study sample, only 11 subjects stated that their mothers had smoked while pregnant with them, but had not smoked in the subjects' childhood. Hence, because of the small size of this group, the effect of prenatal smoking only is difficult to estimate. The effects look stronger when only examining prenatal maternal smoking; this could theoretically be caused by effects among a selected group of women who also smoke later—for instance, women who smoke more.
Second, because the information regarding passive smoking in utero or in childhood was gathered after the incident event, there is a chance that recall bias has influenced the results. For instance, if subjects with incident asthma were more inclined to search for an explanation for their asthma, they may have remembered their mothers smoking to a larger degree than subjects without incident asthma. Thus, the effects could have been overestimated. However, for the effects of passive smoking to have been fully explained by recall bias, one half of subjects with incident asthma would have had to falsely report maternal smoking. Conversely, one half of subjects without incident asthma would have had to falsely underestimate their mothers' smoking. The authors find this unlikely, but caution is advised in interpretation of the findings.
Third, asthma could be subject to misclassification versus chronic obstructive pulmonary disease. However, when we restricted the analyses to subjects younger than 40 years, we still observed a relationship between passive smoking and incidence of asthma. Furthermore, the OR for developing wheezing when exposed to passive smoking was lower than the OR for developing asthma. One could argue that this could be caused by a misclassification of asthma with chronic obstructive pulmonary disease. However, both ORs point in the same direction, and the differences in magnitude of the ORs between asthma and wheezing persisted in subjects younger than 40.
The magnitude of exposure to passive smoking was in line with that reported by other community studies (3, 11). However, the level of exposure varied according to age, personal smoking habits, and educational level. The finding that the oldest subjects barely reported maternal smoking in pregnancy could be caused by a larger tendency for older subjects to forget previous exposure. However, the reported exposures fit well with the dynamic of the tobacco epidemic during the last century. Men took up smoking in large numbers before women, and those with a higher educational level before those with a lower educational level. Thus, maternal smoking in pregnancy was quite uncommon before the 1950s, which fit well with the reported level of exposure. Previous studies have reported that subjects who were exposed to passive smoking in childhood were more likely to take up smoking themselves (13, 27). This finding is supported by the findings in the current study, and serves as a reminder of the importance of early prevention.
Which mechanisms could be responsible for the effects of passive smoking in utero or in childhood on adult incidence of asthma and respiratory symptoms found in this study Although a large body of literature has shown adverse effects of maternal smoking on respiratory health in children, most of the studies cannot separate the effects of pre- and postnatal exposure. However, a growing body of evidence implies that prenatal exposure can be causally related to later obstructive lung disease. Attempts have been made to study breathing patterns in newborns, and three studies have reported a reduction in "tidal breathing ratio" (time to reach tidal peak expiratory flow divided by total expiratory time) in newborn infants of mothers who smoked in pregnancy (28eC30). The clinical significance of this finding in newborns is yet unclear. Several studies conducted on older infants, and which therefore cannot fully separate between pre- and postnatal effects of maternal smoking, have focused on forced expiratory flows at functional residual capacity (maxFRC) (31, 32) and airway resistance (33, 34). Although several studies have found a reduction in maxFRC in infants exposed to smoking in utero (31, 35, 36), the effects on airway responsiveness are yet uncertain (37). Still, several studies have now indicated structural changes in the airways of infants exposed to smoking in utero, and these changes have, at least to some extent, been independent of infant body size. Furthermore, the number of toxic compounds in tobacco is vast, and many will readily pass the placental barrier. Maternal smoking in pregnancy increases fetal carboxyhemoglobin levels and may cause fetal hypoxia (38). Animal studies have suggested that nicotine may disturb alveolar development (39), expression of nicotinic receptors (40), and lung function (41).
The postnatal effects of ETS on children are the result of inhalation of sidestream tobacco smoke. The exact mechanisms are not known. Preschool exposure has consistently been shown to have a greater adverse effect on respiratory health (27). The immature airways may be more sensitive to exposure of passive smoking; however, children spend more time with their mothers in their first years of life and may therefore have a higher level of exposure. The consistent finding that maternal smoking in childhood has a larger effect than smoking by other household members could implicate a doseeCresponse relationship.
The structural or functional changes induced pre- or postnatal by passive smoking could lead to asthma later in life, either through a slow progression of the changes until clinical disease became evident or through inducing vulnerability in the airways for later exposures. Whether factors like sex, atopy, and active smoking influence the effects of passive smoking in childhood is not known. In the Midspan Study, there was evidence that the deleterious effects of current smoking on the ratio of FEV1 to FVC increased with exposure to maternal smoking (12). And, in the report from the European Community Respiratory Health Survey, paternal smoking increased the chance for wheezing among boys, not girls (11). Furthermore, maternal smoking had a larger adverse effect on subjects without atopy (11). In the Hordaland County Cohort Study, the exact opposite trend regarding atopy was observed (data not shown); however, this was not statistically significant. With few studies, and a long time period between the exposure and the outcome, any interactions should be interpreted with caution. Although none were found in the current study, this could be because of sample size, and attempts to examine possible interacting factors should be made in future studies.
In conclusion, the current study implicates that passive smoking in childhood has a lasting effect on the airways, increasing the risk for adult respiratory symptoms and asthma. The preventive potential was found to be quite large, because the fraction of new cases of adult asthma attributable to pre- or postnatal passive smoking was 24%. More studies are needed to verify this finding. Mechanistic studies are needed to further differentiate between pre- and postnatal effects of passive smoking.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
REFERENCES
Beasley R, Crane J, Lai CK, Pearce N. Prevalence and etiology of asthma. J Allergy Clin Immunol 2000;105:S466eCS472.
Brgger J, Bakke P, Eide GE, Johansen B, Andersen A, Gulsvik A. Long-term changes in adult asthma prevalence. Eur Respir J 2003;21:468eC472.
Nafstad P, Kongerud J, Botten G, Hagen JA, Jaakkola JJ. The role of passive smoking in the development of bronchial obstruction during the first 2 years of life. Epidemiology 1997;8:293eC297.
Strachan DP, Cook DG. Health effects of passive smoking: 6. Parental smoking and childhood asthma: longitudinal and case-control studies. Thorax 1998;53:204eC212.
Gilliland FD, Li YF, Peters JM. Effects of maternal smoking during pregnancy and environmental tobacco smoke on asthma and wheezing in children. Am J Respir Crit Care Med 2001;163:429eC436.
Leuenberger P, Schwartz J, Ackermann-Liebrich U, Blaser K, Bolognini G, Bongard JP, Brandli O, Braun P, Bron C, Brutsche M, et al. Passive smoking exposure in adults and chronic respiratory symptoms (SAPALDIA Study): Swiss Study on Air Pollution and Lung Diseases in Adults, SAPALDIA Team. Am J Respir Crit Care Med 1994;150:1222eC1228.
Janson C, Chinn S, Jarvis D, Zock JP, Toren K, Burney P. Effect of passive smoking on respiratory symptoms, bronchial responsiveness, lung function, and total serum IgE in the European Community Respiratory Health Survey: a cross-sectional study. Lancet 2001;358:2103eC2109.
Larsson ML, Loit HM, Meren M, Polluste J, Magnusson A, Larsson K, Lundback B. Passive smoking and respiratory symptoms in the FinEsS Study. Eur Respir J 2003;21:672eC676.
Ng TP, Hui KP, Tan WC. Respiratory symptoms and lung function effects of domestic exposure to tobacco smoke and cooking by gas in non-smoking women in Singapore. J Epidemiol Community Health 1993;47:454eC458.
Thorn J, Brisman J, Toren K. Adult-onset asthma is associated with self-reported mold or environmental tobacco smoke exposures in the home. Allergy 2001;56:287eC292.
Svanes C, Omenaas E, Jarvis D, Chinn S, Gulsvik A, Burney P. Parental smoking in childhood and adult obstructive lung disease: results from the European Community Respiratory Health Survey. Thorax 2004;59:295eC302.
Upton MN, Smith GD, McConnachie A, Hart CL, Watt GC. Maternal and personal cigarette smoking synergize to increase airflow limitation in adults. Am J Respir Crit Care Med 2004;169:479eC487.
Larsson ML, Frisk M, Hallstrom J, Kiviloog J, Lundback B. Environmental tobacco smoke exposure during childhood is associated with increased prevalence of asthma in adults. Chest 2001;120:711eC717.
Duelien T, Eagan TML, Eide GE, Gulsvik A, Bakke PS. Environmental tobacco smoke in childhood and incidence of adult asthma . Eur Respir J 2004;24:(Suppl 48):498s.
Eagan TMM, Duelien T, Eide GE, Gulsvik A, Bakke PS. The attributable fraction of adult asthma and respiratory symptoms due to enviromental tobacco smoke in childhood . Eur Respir J 2004;24(Suppl 48):22s.
Bakke P, Gulsvik A, Lilleng P, Overa O, Hanoa R, Eide GE. Postal survey on airborne occupational exposure and respiratory disorders in Norway: causes and consequences of non-response. J Epidemiol Community Health 1990;44:316eC320.
Eagan TM, Eide GE, Gulsvik A, Bakke PS. Nonresponse in a community cohort study: predictors and consequences for exposure-disease associations. J Clin Epidemiol 2002;55:775eC781.
Eagan TM, Gulsvik A, Eide GE, Bakke PS. The effect of educational level on the incidence of asthma and respiratory symptoms. Respir Med 2004;98:730eC736.
Brogger JC, Bakke PS, Gulsvik A. Comparison of respiratory symptoms questionnaires. Int J Tuberc Lung Dis 2000;4:83eC90.
Gulsvik A, Bakke PS, Humerfelt S, Omenaas E, Baste V. Measurements of respiratory symptoms and sample size to detect a given difference between treatment groups in obstructive lung disease. Eur Respir Rev 1991;1:436eC443.
Rothman KJ, Greenland S. Modern epidemiology, 2nd ed. Philadelphia: Lippincott-Raven; 1998.
Levin ML. The occurence of lung cancer in man. Acta Unio Int Contra Cancrum 1953;9:531eC541.
Whittemore AS. Statistical methods for estimating attributable risk from retrospective data. Stat Med 1982;1:229eC243.
Greenland S, Drescher K. Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics 1993;49:865eC872.
Brady AR. Adjusted population attributable fractions from logistic regression. Stata Technical Bulletin 42 (STB-42). College Station, TX: StataCorp; 1998.
Pocock SJ, Collier TJ, Dandreo KJ, de Stavola BL, Goldman MB, Kalish LA, Kasten LE, McCormack VA. Issues in the reporting of epidemiological studies: a survey of recent practice. BMJ 2004;329:883.
Cook DG, Strachan DP. Health effects of passive smoking-10: Summary of effects of parental smoking on the respiratory health of children and implications for research. Thorax 1999;54:357eC366.
Lodrup Carlsen KC, Jaakkola JJ, Nafstad P, Carlsen KH. In utero exposure to cigarette smoking influences lung function at birth. Eur Respir J 1997;10:1774eC1779.
Stick SM, Burton PR, Gurrin L, Sly PD, LeSouef PN. Effects of maternal smoking during pregnancy and a family history of asthma on respiratory function in newborn infants. Lancet 1996;348:1060eC1064.
Hoo AF, Henschen M, Dezateux C, Costeloe K, Stocks J. Respiratory function among preterm infants whose mothers smoked during pregnancy. Am J Respir Crit Care Med 1998;158:700eC705.
Tager IB, Ngo L, Hanrahan JP. Maternal smoking during pregnancy: effects on lung function during the first 18 months of life. Am J Respir Crit Care Med 1995;152:977eC983.
Young S, Sherrill DL, Arnott J, Diepeveen D, LeSouef PN, Landau LI. Parental factors affecting respiratory function during the first year of life. Pediatr Pulmonol 2000;29:331eC340.
Milner AD, Marsh MJ, Ingram DM, Fox GF, Susiva C. Effects of smoking in pregnancy on neonatal lung function. Arch Dis Child Fetal Neonatal Ed 1999;80:F8eCF14.
Dezateux C, Stocks J, Dundas I, Fletcher ME. Impaired airway function and wheezing in infancy: the influence of maternal smoking and a genetic predisposition to asthma. Am J Respir Crit Care Med 1999;159:403eC410.
Tepper RS, Reister T. Forced expiratory flows and lung volumes in normal infants. Pediatr Pulmonol 1993;15:357eC361.
Adler A, Ngo L, Tosta P, Tager IB. Association of tobacco smoke exposure and respiratory syncitial virus infection with airways reactivity in early childhood. Pediatr Pulmonol 2001;32:418eC427.
Stocks J, Dezateux C. The effect of parental smoking on lung function and development during infancy. Respirology 2003;8:266eC285.
Koren G. Fetal toxicology of environmental tobacco smoke. Curr Opin Pediatr 1995;7:128eC131.
Maritz GS, Dennis H. Maternal nicotine exposure during gestation and lactation interferes with alveolar development in the neonatal lung. Reprod Fertil Dev 1998;10:255eC261.
Sekhon HS, Jia Y, Raab R, Kuryatov A, Pankow JF, Whitsett JA, Lindstrom J, Spindel ER. Prenatal nicotine increases pulmonary alpha7 nicotinic receptor expression and alters fetal lung development in monkeys. J Clin Invest 1999;103:637eC647.
Sekhon HS, Keller JA, Benowitz NL, Spindel ER. Prenatal nicotine exposure alters pulmonary function in newborn rhesus monkeys. Am J Respir Crit Care Med 2001;164:989eC994.