Abstract
The effect of pre- and post-natal smoke exposure on exhaled nitric oxide fraction (FeNO) in infants was evaluated and the association between respiratory symptoms and FeNO in the first 2 months of life was investigated. The Generation R study is a population-based, prenatally recruited birth cohort.
Exposures were assessed by means of questionnaires prospectively administered during pregnancy and after birth. Successful off-line FeNO measurements during tidal breathing were obtained in 187 infants (median age 6.9 weeks). The association between possible determinants and log FeNO was investigated with multiple linear regression analysis.
Infants exposed pre- and post-natally to smoke showed lower FeNO than infants exposed only after birth (geometric mean difference (95% confidence interval) 1.5 (1.0–2.1) ppb) and never-exposed infants (1.4 (1.0–1.8) ppb). FeNO was reduced in infants with severe upper respiratory symptoms compared with infants with nonsevere symptoms (1.6 (1.0–2.4) ppb). Infants with symptoms of the lower respiratory tract had lower FeNO than asymptomatic infants (1.2 (1.0–1.50) ppb).
In conclusion, the nature of the association between smoke exposure and exhaled nitric oxide fraction is dependent on timing and intensity of exposure. The occurrence and the severity of respiratory symptoms in the first 2 months of life are associated with lower exhaled nitric oxide fraction.
- Early respiratory morbidity
- exhaled nitric oxide
- post-natal exposures
- prenatal exposures
- prospective birth cohort
Exhaled nitric oxide fraction (FeNO) is increased in asthmatic adults 1, children 2, and infants with eczema 3 and recurrent wheezing 4, and has been proposed as a noninvasive marker of eosinophilic airway inflammation. Compared with healthy infants, lower FeNO levels have been found in infants with virus-associated acute, wheezing bronchitis 5 and in infants with upper respiratory symptoms (URS), such as rhinorrhoea 6. Several pre- and post-natal factors have been shown to influence the levels of FeNO in infants, such as tobacco smoke exposure 7–9, coffee consumption during pregnancy 8, maternal atopic disease 8, 10, birth weight 11, gestational age 11, 12, sex 8 and infections 13. However, the influence of risk factors for respiratory morbidity on FeNO in infancy is not clear. Previous studies investigating the association between smoke exposure, one of the best known risk factors for respiratory morbidity in infants, and FeNO have given conflicting results. Hall et al. 9 found lower FeNO in infants exposed to smoking during pregnancy than in unexposed infants. In a subsequent report by Frey et al. 8, this difference was only significant in infants of mothers without atopic disease. In addition, the role of post-natal exposure to tobacco smoke in infants has been investigated but the results are not consistent. In a recent study, Franklin et al. 7 reported higher FeNO in infants exposed to post-natal tobacco smoking, whereas previous studies did not show such an effect, or a lower FeNO in exposed asthmatics and healthy subjects 14, 15.
Previous studies that sought to investigate the effect of different determinants of FeNO levels in early infancy retrospectively assessed prenatal exposure variables after birth, rather than prospectively during pregnancy. Therefore, the temporality or succession of events was not documented and the exposure assessment was more prone to recall bias. The aim of the present prospective birth cohort study was to evaluate whether, and to what extent, prenatal and early post-natal exposures influence FeNO in early infancy. The effect of URS and lower respiratory symptoms (LRS) on FeNO in the first 2 months of life was also investigated.
SUBJECTS AND METHODS
The Generation R study is a prospective, population-based, prenatally recruited birth cohort study undertaken in Rotterdam (the Netherlands). A randomly selected group of 1,232 pregnant Dutch females and their children were enrolled in the Generation R focus study. In the focus study, more detailed assessments of foetal and post-natal growth and development were performed 16, 17. Females were enrolled during pregnancy. Sociodemographic factors and exposure to risk factors for respiratory diseases were assessed by means of questionnaires administered to the mother during early (gestational age <18 weeks), mid (gestational age 18–25 weeks) and late (gestational age ≥25 weeks) pregnancy and to the partner at 20 weeks. Information was gathered on the following exposure variables for both the mother and the partner: sociodemographic factors, smoking habits, atopy and siblings. A questionnaire was administered to the parents when the child was 2 months old and exposure variables were again assessed, together with onset, occurrence and severity of URS and LRS.
Between November 2004 and September 2005, FeNO measurements were attempted in 225 infants participating in the focus study at a median (range) age of 6.7 (3.7–16.9) weeks. Mixed oral/nasal FeNO was measured off-line during tidal breathing with a face mask covering the nose and mouth without the use of sedation as previously described 18. An FeNO measurement was considered successful if: exhaled air was sampled during quiet tidal breathing; the face mask was tightly fitted to the nose and mouth during the whole procedure; and at least five breaths were obtained. All FeNO measurements were conducted with the infants awake. Ambient nitric oxide (NO) was measured before each FeNO measurement (Sievers 280 B; Sievers Inc., Boulder, CO, USA). In case of an ambient NO concentration >10 ppb, the infant inhaled ≥2 tidal breaths of NO-free air from an NO-inert 750-mL balloon in order to permit washout of the dead space of the lungs 19. However, as FeNO showed a positive significant association with ambient NO levels, ambient NO was always included in the multivariable models.
All infants were free of respiratory symptoms and had no clinical evidence of airways infection at the time of the measurement. The Medical Ethical Committee of the Erasmus Medical Center (Rotterdam) approved the study. Mothers and their partners received written and verbal information about the study and gave written informed consent.
Definition of variables considered in the analysis
The education level of the mother was divided into three categories (lower, intermediate or higher vocational training) according to the Dutch standard classification of education 20. Parental atopy was defined as self-reported or doctor-diagnosed allergy or atopic disease (allergic asthma, hay fever, eczema). Prenatal maternal smoking was assessed in the first questionnaire by asking the mother whether she had smoked during the pregnancy (no, smoked until the pregnancy was known, or continued smoking after the pregnancy was known). In the second and third questionnaires, the mother was asked whether she had smoked in the past 2 months (no or yes). If the answer was positive at least at two time-points, the infant was classified as prenatally exposed. Exposure to passive tobacco smoking was also assessed by asking whether people regularly smoked in the house or in the working environment of the mother during pregnancy. Gestational age, birthweight and birth length were obtained from the midwife and hospital registries.
Post-natal exposure to tobacco smoke was assessed by asking whether the infant had been exposed to smoke by the mother or by any of the members of the household at least once a week. In addition, parents were asked whether their child had had a runny and/or blocked nose (URS), breathlessness, a whistling noise when breathing, wheezing, panting, difficulty breathing and/or cough (LRS) in the past 2 months. Symptoms were considered severe if they required a visit to a doctor, as reported by the parents.
Statistical analysis
FeNO values were log-normally distributed. Univariable analyses using an unpaired t-test and simple linear regression were used to determine associations between log FeNO and the following explanatory variables: ambient NO; birth weight; gestational age; sex; breastfeeding; maternal educational level; maternal and paternal atopy; pre- and post-natal tobacco smoke exposure; siblings; weight; length at the study date; age at the study date; URS; and LRS. Factors that had a significance level ≤0.1 from univariate analyses were included in the multiple linear regression models in order to evaluate the relationship between log FeNO (dependent variable) and pre and post-natal exposures while controlling for other relevant factors. Although not directly associated with FeNO in the current study population, atopic status of the mother, birth weight, sex and age at the study date were included in all regression models as covariates, as these have been shown to influence FeNO or the occurrence of respiratory symptoms 4, 8. Effect modification by maternal atopy and sex was investigated by adding interaction terms in the final models.
FeNO values were back-transformed after the analysis and are reported as geometric mean (95% confidence interval; CI) in ppb. Comparisons of FeNO between groups are presented as geometric mean of the difference (95% CI). A two-tailed p-value <0.05 was considered significant.
Due to the paucity of data in the literature, no power calculation could be performed in order to evaluate the size of the study needed to detect a difference in FeNO values between groups of infants.
RESULTS
FeNO measurements were attempted in 225 infants and were successful in 187 (83%) infants. During the study, 38 measurements were excluded because a quiet tidal breathing pattern was not maintained during the whole procedure (n = 31) or because <5 breaths were collected in the sampling balloon (n = 7).
Excluded infants had younger mothers (median (range) age 30.3 (18.5–40) yrs) and fathers (32.8 (25.3–39.6) yrs) than infants with successful FeNO measurements (p = 0.003 and p = 0.017, respectively; unpaired t-test), but the other baseline characteristics and anthropometrics at the study date did not differ between the two groups (table 1⇓).
Pre- and post-natal exposures and FeNO
In univariable analysis, anthropometrics were not related to FeNO (table 2⇓), whereas prenatal maternal smoking affected FeNO, with lower levels in exposed infants (p = 0.047; table 3⇓). Paternal smoking and maternal passive tobacco smoke exposure during pregnancy did not affect FeNO; therefore, an infant was considered prenatally exposed if the mother smoked during pregnancy, independent of other sources of smoke. With regard to prenatal smoke, 44 infants were exposed (prenatally only: n = 23; pre- and post-natally: n = 21), whereas 51 infants were exposed to environmental smoke after birth by the mother or by other members of the household (post-natally only: n = 30; pre- and post-natally: n = 21). Sex, parental atopic status and maternal asthma were not related to FeNO (table 3⇓). In order to compare the different smoke exposure categories directly, one variable was created from the combination of pre- and post-natal smoke exposure, with four mutually exclusive categories: never-exposed, prenatal exposure only, post-natal exposure only, and continuous exposure both pre- and post-natally. The association between the combined variable “smoke exposure” and FeNO was significant in the univariable analysis (table 3⇓) and also in the multivariable regression model (table 4⇓). Infants exposed pre- and post-natally to smoke showed lower FeNO than infants exposed only after birth (geometric mean difference (95% CI) 1.5 (1.0–2.1) ppb; p = 0.042) and never-exposed infants (1.4 (1.0–1.8) ppb; p = 0.052; fig. 1⇓). This association was independent of respiratory symptoms and not modified by sex (p for interaction = 0.69) or maternal atopy (p for interaction = 0.46). However, among the 60 infants of atopic mothers, only 22 were exposed to smoke (prenatally: n = 6; post-natally: n = 10; pre- and post-natally: n = 6). The association between ambient NO and FeNO also remained significant (p<0.001) in the multivariable model.
Respiratory symptoms and FeNO
LRS were reported for 83 infants (63 nonsevere and 20 severe), whereas 130 infants had URS (120 nonsevere and 10 severe). In the multivariable analysis, FeNO values were lower in infants with LRS (n = 83) than in asymptomatic (n = 104) infants (geometric mean difference (95% CI) 1.2 (1.0–1.5) ppb; p = 0.046; table 4⇑), but no association was found between the severity of LRS and FeNO. Sex and atopy did not modify this association (p for interaction = 0.68 and 0.88, respectively). Such differences were not found for URS in univariate analysis (p = 0.3), or when URS was added into the multivariable model as a binary outcome (p = 0.9). However, considering the severity of the symptoms (0 = no symptoms; 1 = nonsevere; 2 = severe), infants with a URS that required a doctor’s visit (n = 10) had lower FeNO compared with infants with nonsevere symptoms (geometric mean difference (95% CI) 1.6 (1.0–2.4) ppb; p = 0.047) and tended to have lower FeNO than asymptomatic infants (1.5 (0.9–1.2) ppb; p = 0.075; fig. 2⇓). This effect was independent of smoke exposure and LRS, but still significantly associated with FeNO (table 4⇑). Tobacco smoke exposure or parental atopy, as well as the other investigated determinants, were not associated with occurrence or severity of respiratory symptoms.
DISCUSSION
In the present birth cohort study, an association was found between tobacco smoke exposure and FeNO values in infants. Infants continuously exposed to smoke both in utero and after birth had lower FeNO than never-exposed infants and infants exposed only post-natally. None of the other investigated risk factors for respiratory morbidity affected FeNO. The association between respiratory symptoms and lower FeNO was significant in infants with severe URS and LRS.
Few studies have addressed the separate effects of pre- and post-natal tobacco smoke exposure in infants 7–9, 14. Frey et al. 8 measured FeNO in a selected group of healthy infants and assessed pre- and post-natal environmental tobacco smoke exposure after birth. Frey et al. 8 found that maternal smoking in pregnancy was associated with lower FeNO, but only in infants of mothers without asthma, whereas the same exposure in mothers with atopic disease was associated with higher levels. In univariable analysis, it was found that FeNO was lower in infants of mothers who smoked during pregnancy, whereas in the multivariable analysis infants exposed to smoke prenatally had lower FeNO than never-exposed infants only if smoke exposure was protracted after birth. In the present study, neither parental atopy nor sex modified this association, confirming the earlier findings of Hall et al. 9. However, due to the small numbers of infants per group, there was insufficient power to adequately investigate such interactions. The current study is embedded in a larger population-based birth cohort, and infants were not selected according to their health status. Furthermore, the repeated assessment of exposure variables during pregnancy provided an opportunity to study prenatal smoke exposure in greater detail and reduced the likelihood of recall bias, strengthening the validity of the current findings. In a recent study, Franklin et al. 7 found increased FeNO in infants exposed to parental smoking and evidence of a dose–response relationship. In the present study, a clear effect of post-natal smoke exposure on FeNO could not be demonstrated, as it was found that FeNO was higher in infants exposed to smoke only post-natally than in infants exposed both pre- and post-natally, but it was not different from never-exposed infants. The mechanisms for increased FeNO in infants exposed to post-natal maternal smoking, as observed previously and herein, are not clear. A possible explanation may be a direct irritant effect of smoke on the airways 7.
The lower FeNO found in infants exposed to prenatal tobacco smoking would support the hypothesis that smoke exposure during pregnancy inhibits inducible NO synthase 21. Possible implications of such suppression are hypothetical, but as NO may serve important functions in local defence and in the maintenance of normal vaso- and bronchomotor tone, any factor that modifies baseline NO generation in the airways of young infants should be a reason for concern and further study.
The occurrence and severity of respiratory symptoms were associated with lower FeNO in infants. Franklin et al. 6 also found low FeNO in infants with ongoing rhinorrhoea, but FeNO increased 4–12 weeks after the initial assessment, when symptoms had resolved. Although children in the present study were free of respiratory symptoms and had no evidence of respiratory infection at the time of testing, they might have had symptoms in the weeks preceding the FeNO measurement. Therefore, the current results should be interpreted with caution, as the age at which symptoms occurred and the timing of symptoms in relation to the FeNO measurements might have influenced the findings. It cannot be excluded that a reduced FeNO in infants with respiratory symptoms is related to a delayed effect of acute symptoms on the NO generation or diffusion through the airways. Another reason for caution, when interpreting the present results, is that this association was shown when FeNO was compared between the relatively small group of infants with severe URS and infants with nonsevere symptoms. A child may be taken to the doctor for many reasons, such as symptom severity, but parental anxiety could also have influenced the decision, and this might have led to misclassification. However, if such misclassification occurred, this is not likely to bias the direction of the current results as this would mean that the effect size was underestimated.
Lower FeNO was found in infants who had had LRS, but no association was found between wheezing and FeNO. Previous studies found an association between wheezing and FeNO in selected populations of infants with a high risk of developing atopic disease or with recurrent wheeze 4, 22. In addition, in a prospective study 10, higher FeNO at 1 month of age was predictive of the development of respiratory symptoms in the first year in infants of atopic mothers only, who are at higher risk of developing asthma. In contrast, in the same study, a trend towards a negative association between FeNO and severe respiratory symptoms in infants of nonatopic mothers was found. Infants in the present study were from an unselected population and maternal atopy did not modify any of the associations found, which may explain the discrepancies. Although there is evidence that most asthma starts early in life 23, respiratory morbidity in the preschool child is mostly related to neutrophilic airway inflammation 24 and there is very little evidence of chronic eosinophilic bronchial inflammation in the first months of life 25.
It could be argued that measuring mixed oral/nasal FeNO without controlling for expiratory flow might introduce variability 26, especially in infants exposed to tobacco smoke who may have abnormal airway mechanics. It has previously been demonstrated that FeNO measured with variable flow was reproducible 18 and allowed differentiation between infants with different respiratory diseases in a similar way to more sophisticated techniques, taking into account lung function parameters and breathing pattern 4. However, differences between groups may still be due to differences in tidal flows, particularly when comparing groups with potentially different tidal breathing patterns. Indeed, differences in tidal flows might introduce variability and explain some of the overlap and the relatively small differences between the groups in the present study.
The current authors found a positive correlation between FeNO and ambient NO, in agreement with a previous study by Pijnenburg et al. 27. Although the correlation was also significant in the multivariable model, the current authors showed that it did not affect the results of the study, as the associations found were independent of ambient NO concentrations. In infants, the influence of ambient NO on FeNO could be reduced by always using NO-free air when measuring FeNO 28. However, such recommendations were not available when the present study commenced, and NO-free air was used only if room concentrations were >10 ppb, in agreement with previously published guidelines 19. The current findings suggest that infants should inhale >2 tidal breaths of NO-free air, as this would reduce the contamination by ambient NO; however, this might reduce the success rate of the measurements, as awake infants might not tolerate the face mask for a longer period of time.
A possible limitation to the present study is that smoke exposure was assessed by means of questionnaires and not confirmed with the measurement of specific biomarkers. Although a good agreement between parental report of smoking and air nicotine concentration has been shown 29, some misclassification might have occurred, as parents would under-report smoking. If this was the case, the effect of smoke on FeNO would be underestimated, as smoking parents would be classified in the group of nonsmokers. Therefore, the current authors hypothesise that the size of this effect could be greater than reported.
In conclusion, pre- and post-natal tobacco smoke exposure are associated with mixed oral/nasal exhaled nitric oxide fraction in early infancy, with lower exhaled nitric oxide fraction in prenatally exposed infants and higher exhaled nitric oxide fraction in case of post-natal exposure. Reported airway symptoms, depending on the frequency and severity, were associated with lower exhaled nitric oxide fraction in the first 2 months of life. The meaning of changes in exhaled nitric oxide fraction for respiratory health in infancy needs to be further elucidated.
Support statement
The first phase of the Generation R study received financial support from: Erasmus Medical Center (Rotterdam, the Netherlands), Erasmus University (Rotterdam) and the Netherlands Organization for Health Research and Development (ZonMw). The present study was supported by an additional grant from the Netherlands Asthma Foundation (project number 3.2.02.41).
Statement of interest
A statement of interest for J.C. de Jongste can be found at www.erj.ersjournals.com/misc/statements.shtml
Acknowledgments
The authors are thankful to C. Wagemakers, K. van Willigen-Broekhuize and M. van Leeuwen (Dept of Paediatrics, Sophia Children's Hospital, Rotterdam, the Netherlands) for assistance in sampling FeNO in infants. The authors also wish to thank W.C. Hop (Dept of Epidemidogy and Biostatistics, Erasmus University Medical Center, Rotterdam, the Netherlands) for statistical advice.
The Generation R study was conducted by the Erasmus Medical Center in close collaboration with the School of Law and Faculty of Social Sciences, Erasmus University (Rotterdam, the Netherlands), the Municipal Health Service (Rotterdam), the Rotterdam Homecare Foundation and the Stichting Trombosedienst en Artsenlaboratorium Rijnmond (STAR; Rotterdam). The authors gratefully acknowledge the contribution of general practitioners, hospitals, midwives and pharmacies in Rotterdam.
Footnotes
-
For editorial comments see page 252.
- Received October 9, 2007.
- Accepted March 20, 2008.
- © ERS Journals Ltd