In Utero and Postnatal Maternal Smoking and Asthma in Adolescence.

Abstract


Background:

Asthma in early childhood has been associated with maternal smoking during pregnancy and parental smoking soon after birth. However, less is known about these exposures and the development of asthma symptoms in adolescence.

Methods:

Data were taken from the Mater University Study of Pregnancy, a large birth cohort study of mothers and children enrolled in Brisbane, Australia, beginning in 1981. Smoking was assessed at 2 stages during pregnancy and at the 6-month and 5-year follow-up visits. Asthma was assessed from maternal reports that were provided when the child was age 14 years. We conducted multivariable multinomial logistic regression analyses to assess the effect of maternal smoking on asthma symptoms.

Results:

There was a strong sex interaction such that girls whose mothers had smoked heavily (20 or more cigarettes per day) in pregnancy and at the 6-month follow up had increased odds of experiencing asthma symptoms at age 14. The contribution of heavy smoking during pregnancy appeared to be stronger than heavy smoking after the birth. No similar associations were seen for boys.

Conclusion:

Female adolescents whose mothers smoked heavily during the fetal period and the early months of life have increased risk of asthma symptoms in adolescence. In utero exposure to heavy smoking was found to have a stronger effect than postnatal environmental tobacco exposure.

Introduction

There is considerable evidence linking parental environmental tobacco smoke in early infancy with the development of asthma and wheezing illness in early childhood. A meta-analysis suggests that children exposed to tobacco smoke are at twice the risk of developing serious respiratory problems compared with the children of a nonsmoking parent, with the risk being greater if the parents smoke when the child is under the age of 2 years. This is especially true if the environmental exposure is a consequence of maternal smoking and, in some studies, a dose–response association has been observed. Evidence from animal studies also suggests that environmental tobacco smoke results in reduced lung volume, lung hyperplasia, and reduced gas-exchange surface.

Maternal smoking in the fetal period may also contribute to the development of infant onset of respiratory problems. In 1 study, the strength of the association was only slightly reduced by adjustments for the mediating effects of low birth weight and gestational age. In other studies, retrospective assessments of maternal smoking at different points in time have found that the magnitude of the association between in utero tobacco smoke exposure on childhood asthma was greater than that of postnatal exposure. This evidence has led to the hypothesis that cigarette smoke exposure in pregnancy may have a longlasting impact on the fetus’ airways by reducing the rate of fetal breathing movements and predisposing to limited airway patency in infancy 9 and reduced size of placental arteries.

The association of maternal tobacco use in the prenatal and immediate postnatal period with asthma persistence beyond childhood is, however, less well established. There are only a handful of longitudinal studies with the capacity to assess both maternal smoking at either (or both) of these periods and asthma in adolescence. Findings from the Dunedin longitudinal study reported the effect of postnatal exposure only and suggested an association between parental smoking and airflow reduction at age 15, but only in boys. In contrast, results from the 1958 British birth cohort assessed maternal smoking in pregnancy only and found that it was associated with wheezing from 0 to 7 years. However, the odds of these children developing asthma declined at the ages of 7 and 16.

A similar decrease in risk was found among children age 5 and 16 years participating in the 1970 British birth cohort study. Follow up of the British and New Zealand cohorts into young adulthood found that active smoking of the child became a much stronger predictor of persistent asthma symptoms or relapse at ages 26 and 33 years. Therefore, it is uncertain whether maternal tobacco use in utero and immediately after birth continues to exert an effect on adolescent asthma. Our study examined associations of maternal smoking during pregnancy, after birth, and during early infancy with asthma reported at the age of 14 years in a prospective birth cohort study of Australian children.

DISCUSSION

These results extend the existing evidence on a relationship between maternal smoking and childhood asthma to suggest a persistence of the effect among female adolescents. In our study, heavy maternal tobacco consumption at specific points in time increased the odds of developing asthma among adolescent girls. When we explored the contribution of smoking in each of these periods, we found that the risk of developing asthma was increased at 3 particular points in time: in early and late pregnancy and 6 months after the birth. Earlier exposure (when mothers may have not been aware of being pregnant) and exposure in later childhood were not associated with increased prevalence of symptoms. Consistent with another study, we found that the risk became apparent only when mothers smoked heavily.

Girls who were prenatally exposed to maternal smoking of 20+ cigarettes a day were at an increased risk of asthma at age 14 when compared with those exposed later, but not during pregnancy. These results support evidence from the Children's Health Study of a causal association between tobacco exposure specific to the prenatal period and the development of asthma. It adds to these findings by suggesting that the association extends into adolescence, although only for girls. Our reports of maternal smoking were obtained at the time of pregnancy, unlike mothers’ retrospective recall of smoking during pregnancy available from the Children's Health Study. We had multiple prospective assessments of both maternal in utero and postnatal tobacco use.

Our data on intensity and duration of the exposure provide evidence for a dose–response effect as reported by Gilliland and colleagues. It is important to note that, like others, we could not totally distinguish those who smoked heavily only during pregnancy from those who smoked heavily only after the child's birth. Mothers who smoked 20 cigarettes or more per day during pregnancy were also likely to maintain this level of smoking after the child's birth. Because it was difficult to completely disentangle the relative contribution of each period, our findings may be interpreted as pointing to an accumulative risk from a persisting exposure to heavy smoking over the pre- and postnatal period. However, the fact that we successfully separated those who had not been exposed to heavy smoking during pregnancy and that this group was at a lower risk of having asthma at age 14 suggests that maternal heavy tobacco use during pregnancy may represent the most sensitive exposure.

To examine the impact of cumulative exposure on asthma symptoms in girls, we explored the effect in those who reported consistently smoking 20 cigarettes or more per day in pregnancy and at the 6-month follow up. In this analysis, the cumulative effect of this level of maternal smoking only mildly increased the risk of current asthma symptoms in girls when compared with the effect of only prenatal exposure. In addition, factors that may confound the association between in utero heavy tobacco exposure and asthma symptoms (hereditary condition, early onset of asthma, child's own smoking, and paternal tobacco exposure) did not attenuate the strength of the associations in our findings, suggesting that in utero tobacco exposure may come first in the causal pathway to asthma among adolescent girls. Further studies with greater capacity to separate smoking exposure at these different points in time are needed to replicate our findings and to further our understanding of the specific contribution of in utero tobacco exposure to the development of asthma in adolescent girls.

We consistently found a sex difference such that the association between early smoking exposures and asthma increases the risk of asthma only for girls. This does not confirm findings from previous longitudinal evidence, which found an association between parental smoking (after birth only) and asthma only in adolescent boys. Because other studies did not find the same sex difference, our results are difficult to interpret. Our sample was assessed at age 14 years, which is an age during which boys and girls undergo different stages of pubertal development. By this time, most girls have completed their pubertal growth spurt, whereas boys experience a slower rate of growth, which lasts for a longer period of time. It may be that smoking exposure in utero exerts more serious damage to airway function in females, and this may be explained by biologic differences in lung size between males and females.

In addition, the interplay between airway function and sex hormones at pubertal age may affect the prevalence of asthma at this particular time. We examined the possible impact of puberty on the relationship between maternal smoking and asthma on a subsample of our female cohort (n = 1402) for which we had information on whether girls had experienced the first menstruation at the time of the age-14 assessment. We did not find an association between any level of maternal smoking and asthma among the 263 girls who had been assessed before the onset of menses. However, in girls who had been assessed after the onset of menses, we found the same associations as in our general sample between in utero and postnatal tobacco exposure and asthma at age 14 years.

Interaction tests did not show a statistical difference between the 2 menses-status groups, which may reflect the small number of cases among those who had been assessed before the onset of menses. Further studies with the capacity to test this hypothesis are needed, as well as studies on older cohorts to determine whether these levels of maternal in utero smoking exposure continue to exert an effect on female asthma later in adulthood.

Our results should be interpreted in the context of some limitations, the most notable of which is the reliance on maternal reports of asthma and smoking. Although we used 2 maternal reports for our asthma definition, the fact remains that these reports do not constitute a clinical diagnosis of asthma. Our smoking measures were based on maternal self-reports and not validated by objective measurements of cotinine levels. The study had poor measures of environmental tobacco smoke. Although we used a proxy measure of environmental tobacco smoke to control for paternal tobacco exposure at age 14, we had no data on intensity and duration of paternal tobacco exposure, smoking by others in the household, or other possible environmental factors such as environmental dust pollution, which are found to impact on the association between environmental smoking exposure and asthma later in life.

Another limitation of the study is the level of attrition of those lost to follow up, which may bias our findings. Accounting for the bias resulting from nonresponse is important in the analysis of longitudinal data. If the associations reported here were less prevalent among nonrespondents, our results would be overestimating the association between maternal smoking and reports of asthma at age 14. However, because we typically find that mothers lost to follow up in our study are more likely to be smokers and their children are more likely to exhibit a range of health problems, this seems unlikely. To further assess whether those lost to follow up produced bias in our results, we attached inverse probability weighting to subjects included in the analyses to restore the representation of those lost to follow up. We followed the method suggested by Hogan and colleagues with robust standard errors estimates applied to the model. We found little difference between the weighted and unweighted results (data not presented), which suggests that attrition is unlikely to have substantively biased our findings.

Despite these limitations, this study has the strength of being longitudinal and one of the few studies to prospectively examine the association of maternal smoking at several points in time and asthma symptoms in adolescence, taking account of a number of important potential confounding factors. Our findings provide preliminary evidence of a causal association between in utero maternal heavy smoking and adolescent asthma in girls. Other longitudinal data with clinical assessments of respiratory functions are needed to replicate the gender difference by pubertal age.

In conclusion, although maternal smoking before recognition of pregnant status or in early childhood was not associated with increased risk of asthma symptoms at age 14, heavy maternal smoking during and immediately after pregnancy increased the risk of asthma among female adolescents.