Our results indicate that maternal smoking during pregnancy is strongly associated with persistent asthma in our sample of black, Puerto Rican, and Mexican pediatric asthma patients. Odds of in utero exposure to tobacco smoke were 3.6 times as likely among persistent asthmatics than among intermittent asthmatics. In utero smoking was also associated with daily and nocturnal symptoms as well as asthma-related emergency department visits, and these differences remained after controlling for a number of known risk factors, including current tobacco smoke exposure, ethnicity, plasma IgE, and atopic history. Furthermore, early life tobacco smoke exposure had no effect on the association between in utero tobacco smoke exposure and asthma severity.
In the United States, 13.8% of women in 2005 reported smoking during pregnancy.30
Although the prevalence of maternal smoking during pregnancy has consistently declined during recent years, the 2005 estimate reflects a change of <2 percentage points from 2000.30,31
More recently, data from 2008 and 2009 showed that the 1-month prevalence of cigarette smoking among pregnant women was 15.3% as compared with 27.4% for nonpregnant women.32
The high prevalence of maternal smoking during pregnancy, along with the reported 2009 overall rates of cigarette use of 19.2% for black women and 9.8% for Hispanic women,33
is an issue of concern for the numerous individuals with increased health risks as a result of this preventable exposure.
Maternal smoking during pregnancy is known to vary according to race/ethnicity. In 2007, the percentage of mothers in the United States who smoked while pregnant was lower for Mexican Americans (1.7%) than for Puerto Ricans (7.4%) or black mothers (7.7%).34
These differences in in utero exposure are consistent with the observation that asthma morbidity and mortality are lowest among Mexicans and greatest among Puerto Ricans and blacks.35
Our results indicate that regardless of ethnicity, smoking while pregnant is associated with increased asthma morbidity.
The strong association between tobacco smoke exposure and persistent asthma seems to be seeded during the relatively short 40-week period of pregnancy, underscoring the importance of tobacco prevention, control, and cessation efforts. Smoking bans have been shown in systematic reviews to increase smoking cessation36
and decrease exposure to secondhand smoke (SHS).37
A number of city and county governments have recently passed successful initiatives to reduce SHS in public outdoor areas. As of October 2010, an estimated 79% of the US population was protected by smoke-free ordinances covering workplaces, restaurants, and/or bars.38
The passage of these initiatives is a success for tobacco control and is a stimulus that may lead to more smoke-free homes.39
However, the efficacy of outdoor smoke-free laws seems to be muted among children with at-home SHS exposures,40
particularly because the majority of children's exposure to SHS occurs in homes.41,42
Although cessation of smoking before, during, and after pregnancy is the preferred ideal, our results reveal that abstaining from smoking during pregnancy may significantly reduce the burden of persistent asthma. Nine months of abstinence may also be more acceptable to those who are not considering smoking cessation, but care should be taken to avoid conveying the false impression that abstaining from smoking during pregnancy is just as safe as not smoking at all.
Our findings are tempered by several limitations. Although a prospective study design (eg, birth cohort study) likely could have produced conclusions more robust to certain biases, our study is, to our knowledge, the first in which the effects of in utero tobacco smoke exposure on asthma severity and exacerbations among minority populations have been investigated. Novelty notwithstanding, the retrospective nature of our case-control study design may have introduced biases in ascertainment of exposures. Given that data regarding tobacco smoke exposure were collected retrospectively, tobacco smoke exposure was not confirmed via objective measures such as the measurement of cotinine, a sensitive measure of tobacco smoke exposure. There is also the potential that parents of persistent asthmatics might have systematically differed in the way they recalled important exposures (eg, smoking during pregnancy) compared with parents of intermittent asthmatics. Respondents of children with more severe disease, for example, may have been influenced by remorse or anxiety to report more or less exposure. Respondents may also have been motivated to give socially desirable responses (ie, to underreport tobacco smoking). However, the reported prevalence of smoking during pregnancy in our study was comparable to national estimates (9.2% in our study; 13.8% nationally30
), and the proportion of in utero tobacco smoke exposure in our study population trended in the expected direction (ie, greater among persistent asthmatics than among intermittent asthmatics).
We acknowledge that exposures to tobacco smoke not measured in our study (eg, smoking by nonparent household contacts) also could have affected asthma severity and exacerbations. However, 90% of tobacco smoke exposure in children occurs from parents.43
Given that Latino families are more likely to have intergenerational households compared with white and black families,44
it is possible that additional sources of environmental tobacco smoke exposure in our Latino population could have come from caregivers and other relatives living in the home. Nonetheless, in a community-based study of inner-city children with asthma (82% of which were Hispanic or black), 94% of caregivers were parents, and smoking status of these caregivers was more strongly associated with increased cotinine levels in children than smoking status of other household members.45
On the basis of these observations, we assumed in our analyses that parental exposures are generally more common and influential than nonparental exposures for the majority of our study population. Likewise, this supports our use of parental smoking status as a proxy for tobacco smoke exposure in the child.
An additional limitation of our study was that we did not quantify the amount or duration of tobacco smoke exposure, which might have been helpful for estimating dose-dependent effects. Although we did not measure exposures to other sensitizing agents (eg, major indoor allergens), we included terms for serum IgE and atopic dermatitis in regression models as proxies for measuring atopic status in our study population. Accounting for the individuals with atopic status is important in light of the accepted observation that atopic individuals have an increased prevalence of asthmatic symptoms, including wheezing, and hospital visits.46
In addition, the generalizability of our findings may be limited because we studied a specific subset of the population (black, Puerto Rican, and Mexican). However, our results are directly pertinent to the segment of the population most affected by asthma.
Our study also has a number of strengths. Recruitment of participants was subject to strict inclusion and exclusion criteria to minimize information biases. Ascertainment bias of disease status is unlikely, as asthmatic patients in our study were physician-diagnosed and well phenotyped, including by spirometry. Misclassification of race/ethnicity is also unlikely because 1 inclusion criterion required that all biological parents and grandparents identify as the same race/ethnicity. Our use of a multiethnic study population is another strong point of our study, given that the majority of previously published studies have been conducted in white/European populations.
Sixteen years have passed since tobacco use was declared a pediatric disease by the commissioner of the US Food and Drug Administration in 1995.47
Yet today, a significant proportion of children still remain exposed to SHS. What can clinicians do to help break the perpetuation of health consequences of smoking? The American Academy of Pediatrics policy statement on tobacco use provides recommendations and guidelines for physicians that cover personal and professional behavior, clinical practice, and advocacy.48
Clinicians are recommended, in part, to ask patients about tobacco use during each clinical visit. This screening has been identified by the US Public Health Service as the most important step in addressing tobacco use and dependence.49
These recommendations are relevant given that compared with non-Hispanic whites, fewer Hispanic patients are offered screening and counseling for tobacco use during primary care visits,50
and black and Latino patients are both less likely to receive tobacco cessation interventions51
or use nicotine replacement therapy for smoking cessation.52