Parental smoking and development of allergic sensitization from birth to adolescence.

BACKGROUND: The relation between secondhand tobacco smoke (SHS) exposure and the development of allergic sensitization in children is unclear. The aim of this study was to determine whether maternal smoking during pregnancy and postnatal SHS exposure contributes to the development of allergic sensitization in children and adolescents up to 16 years of age.
METHODS: We included 3316 children from a birth cohort followed up for 16 years. SHS exposure and symptoms of allergic disease were assessed using repeated parental questionnaires. Serum immunoglobulin E against eight common inhalant and six food allergens was assessed at ages 4, 8, and 16 years with ImmunoCAP. The association between SHS exposure and sensitization was explored using logistic regression and generalized estimating equations.
RESULTS: Exposure to SHS in infancy without prior exposure in utero was associated with an excess risk of food sensitization at age 4 years (OR 1.47, 95% CI 1.08-2.00), with comparable ORs at ages 8 and 16 years. In longitudinal analyses, an overall association was indicated between SHS in infancy and food sensitization up to age 16 years (OR 1.24, 95% CI 0.98-1.56). Maternal smoking during pregnancy was unrelated to sensitization up to 16 years of age. When sensitization was combined with concurrent symptoms of allergic disease, SHS in infancy was associated with an overall elevated risk of eczema with sensitization (OR 1.62, 95% CI 1.20-2.18).
CONCLUSIONS: SHS exposure in infancy appears to increase the risk of sensitization to food allergens up to age 16 years, as well as eczema in combination with sensitization.

The prevalence of atopic disease has increased over recent decades, and environmental risk factors and gene‐environment interactions play a probable role 1, 2. Despite a declining prevalence of cigarette smoking in many countries, secondhand tobacco smoke (SHS) remains a pervasive indoor environmental risk factor in children of smoking parents. SHS has immunomodulatory effects and is hypothesized to influence the development of allergic sensitization 3. SHS exposure during the fetal period and thereafter has been associated with both acute and long term adverse health consequences in children and adolescents 4. Mounting evidence links exposure to SHS with respiratory and allergic disease in children, but the impact of tobacco smoke on immunoglobulin E (IgE) mediated sensitization is less clear 5. Some studies report increased risks for any allergen sensitization 1, 6, whereas others report increased risks only for food allergen sensitization 7, 8. Studies have also reported inverse 3, 9, 10 or null associations for inhalant allergens 4 and there are conflicting data on the role of heredity 1, 3, 6. Possible explanations for the difference in observed results include insufficient control of confounding, the grouping of inhalant and food allergens together, failure to consider possible differing effects of SHS between those with and without family history of atopy, or maternal vs paternal smoking 3. Teasing apart the effects of in utero and postnatal SHS exposure remain challenging, but important nonetheless, as the etiological mechanisms may differ depending on timing of exposure.

Few studies have followed children longitudinally from birth to adolescence, and even fewer have collected multiple blood samples. Due to the lack of consistent evidence, we sought to evaluate whether pre‐ and postnatal SHS exposure contributes to the development of allergic sensitization in a population‐based birth cohort study followed longitudinally through adolescence.

Methods

Study subjects

The BAMSE (Barn/Child, Allergy, Milieu, Stockholm, Epidemiology) study is a longitudinal population‐based birth cohort in which infants were recruited at birth and prospectively followed during childhood and adolescence. A total of 4089 infants born in Stockholm, Sweden between 1994 and 1996 were included 11. At a median infant age of 2 months, parents completed the baseline questionnaire which assessed environmental exposures, parental smoking habits, residential characteristics, lifestyle, and parental allergies. When children were approximately ages 1, 2, 4, 8, 12, and 16 years, parents completed questionnaires focusing on symptoms of asthma, rhinitis, and eczema in their children and current parental smoking habits 12. Response rates from baseline were 96%, 94%, 91%, 84%, 82%, and 78%, respectively.

Blood samples were collected at ages 4, 8, and 16 years, and serological allergy IgE testing was performed in 2605 (63.7%), 2470 (60.4%), and 2547 (62.2%) children, respectively.

The baseline and all follow‐up studies were approved by the Regional Ethical Review Board, Karolinska Institutet, Stockholm, Sweden, and the parents of all participating children provided informed consent.

Secondhand smoke exposure assessment

Fetal exposure to maternal smoking was defined as daily maternal smoking of one or more cigarettes at any time during pregnancy. Information on maternal smoking during pregnancy was collected at baseline when children were a median age of 2 months. SHS exposure in infancy was defined as maternal and/or paternal smoking of one or more cigarettes per day at baseline. SHS at ages 1, 2, 4, 8, 12, and 16 years of age were defined as maternal and/or paternal smoking of one or more cigarettes daily at the time of the respective follow‐up. Information on the average number of cigarettes smoked daily by each parent was collected at each follow‐up.

Assessment of allergic sensitization

Allergic sensitization was defined based on IgE antibody reactivity against common inhalant and food allergens. At each follow‐up, sera were screened with Phadiatop®, a mix of typical inhalant allergens: pollens of birch, timothy, and mugwort, danders of cat, dog and horse, mold [Cladosporium herbarum], and house dust mite [Dermatophagoides pteronyssinus], and fx5®, a mix of typical food allergens: cow's milk, hen's egg, soy bean, peanut, cod fish and wheat (ImmunoCAP System; Thermo Fisher/Phadia AB, Uppsala, Sweden). Sera that scored positive for Phadiatop® or fx5® were subsequently analyzed for allergen‐specific IgE antibodies to the single inhalant and food allergens mentioned above. Any allergen sensitization was defined as a positive reaction, 0.35 kUA/l or more, to any of the allergens tested. Inhalant allergens were classified as either indoor (cat, dog, horse, and/or house dust mite) or outdoor (timothy grass, birch, mugwort, and/or mold). The reference groups were made up of subjects not sensitized against the allergens being tested.

Assessment of asthma, rhinitis and eczema

The assessment of asthma, rhinitis and eczema were based on parental questionnaires at ages 4, 8, and 16 years.

Asthma: four or more episodes of wheeze in the last 12 months or one or more episode of wheeze in the last 12 months in combination with inhaled steroids 12.

Rhinitis: eye or nose symptoms following exposure to allergens in the last 12 months and/or a doctor's diagnosis of allergic rhinitis 12.

Eczema: dry skin, pruritic rash with age‐specific localizations at face, flexures of arms or legs, wrists or ankles, or neck in the last 12 months and/or a doctor's diagnosis of eczema 12.

Statistical analyses

Chi‐square tests were used to test for differences in covariates between groups of children exposed to maternal smoking during pregnancy (in utero) and in infancy. Associations between SHS exposure and allergic sensitization were analyzed with SHS allocated into four categories: (i) no exposure in utero or in infancy (reference category), (ii) only in utero, (iii) only in infancy, and (iv) both during in utero and in infancy. The association between these categories and sensitization were analyzed using multinomial logistic regression. All associations are reported as odds ratios (ORs) and 95% confidence intervals (CIs), and P < 0.05 was considered statistically significant. Associations between SHS and allergic sensitization were also analyzed longitudinally using generalized estimating equation (GEE) models with an unstructured correlation matrix 13. This model included an interaction term between the time indicator variable and exposure to evaluate the effect of exposure over time 13. In these analyses, exposure to maternal smoking during pregnancy and parental smoking in infancy were dichotomized (yes vs no) to increase power and robustness. To examine dose–response associations between the numbers of cigarettes smoked per day and sensitization, we defined three categories for exposure to maternal smoking during pregnancy: (i) no cigarettes throughout pregnancy (reference category), (ii) 1–9 cigarettes per day during any trimester, and (iii) ≥10 cigarettes per day during any trimester. Similar categories were coded for parental smoking in infancy, for each parent individually and then combined: (i) mother and father did not smoke (reference category), (ii) mother/father smoked 1–9 cigarettes per day, and (iii) mother/father smoked ≥10 cigarettes per day.

To account for the effect of smoking throughout early childhood and adolescence a ‘smoking throughout childhood’ variable was constructed. This was defined based on the responses (yes vs no) in the questionnaires prior to the blood sample collection at 4, 8, and 16 years. For example, parental smoking assessed at ages 1 and 2 were combined and regressed against outcomes assessed at year 4.

To elucidate the effect of SHS on sensitization combined with symptoms of asthma, rhinitis and eczema, categorical variables were created. Asthma was categorized into: (i) no asthma and no sensitization (reference category), (ii) asthma without any sensitization, (iii) asthma with any sensitization, and (iv) any sensitization only. Similar categories were created for rhinitis and eczema. Corresponding categories were also used in analyses focused on food allergen sensitization and symptoms of asthma, rhinitis, and eczema.

Various covariates were tested through exploratory stepwise logistic regression to identify potential confounding factors. Models tested included sex, duration of breastfeeding, socioeconomic status (categorized on the basis of parents’ occupation as manual and nonmanual workers), construction year of home, presence of siblings, maternal age, reported dampness or mold in the home at birth, air pollution from local road traffic 14 (using NOx as indicator), parental allergic disease, cat or dog ownership, birth weight and children's own smoking history. Covariates leading to > 5% change in estimates were kept in the final models and included parental allergy and socioeconomic status. To disentangle the effects of maternal smoking during pregnancy and parental smoking in infancy from that of exposure to SHS at ages 1–16 years, we created models adjusted for SHS throughout childhood (ages 1–16 years).

Because GEE models can provide an estimate when observations are missing or unequally spaced, we included all children who gave at least one blood sample in the final analyses 13. A total of 3316 children with complete information on exposure, confounders, and blood samples were included in the final analyses.

All statistical analyses were performed with STATA (release 12; Stata Corp., College Station, TX, USA).

Results

The 3316 children included in this study were comparable to the original cohort regarding distribution of background characteristics such as SHS exposure (Table S1). In total, 12.6% were exposed to maternal smoking during pregnancy and 20.8% to SHS by any parent in infancy. A substantial proportion of mothers who smoked in infancy had also smoked during pregnancy (Figure S1). Maternal smoking during pregnancy, low parental socioeconomic status, low maternal age, and furred pets at home were more prevalent among children exposed to parental smoking in infancy (Table S2). Parental allergy and longer duration of breastfeeding were less frequent among children exposed to SHS in infancy. Similar patterns were observed in children exposed to maternal smoking during pregnancy (data not shown). Parental smoking was highest in infancy and declined to 13.5% at the 16 year follow‐up.

Prevalence rates of sensitization to any inhalant or food allergen at ages 4, 8, and 16 were 24.1%, 34.8%, and 45.9%, respectively (Table S3). There was considerable overlap between any sensitization at ages 4, 8, and 16 years, indicating that few children lost their sensitization (Fig. 1) 15. The prevalence of asthma, rhinitis, and eczema at age 16 years were 6%, 26%, and 9%, respectively.

Figure 1

Prevalence and proportional relationship of any sensitization at 4, 8, and 16 years among children in the BAMSE birth cohort (n = 1682).

To try to disentangle the influence of pre‐ and postnatal SHS exposure, maternal smoking during pregnancy and parental smoking in infancy were allocated into four categories. Among children exposed to maternal smoking during pregnancy but not SHS in infancy, we observed no association with sensitization to any (inhalant or food) allergen at ages 4, 8, and 16 years (Table 1). However, children exposed in infancy without prior exposure to maternal smoking during pregnancy, had an increased risk of food sensitization at age 4 years (OR 1.47, 95% CI 1.08–2.00), and a tendency for increased risks which were no longer statistically significant at 8 (OR 1.34, 95% CI 0.99–1.79) and 16 years (OR 1.32, 95% CI 0.93–1.86). Moreover, these children had significantly increased odds of persistent food allergen sensitization up to age 16 years (OR 2.17, 95% CI 1.32–3.56) (Table S4). In addition, there was an increased risk of sensitization to indoor allergens at age 4 years for these children (OR 1.50, 95% CI 1.00–2.24). We observed no significantly increased risk of any sensitization when children were exposed to both maternal smoking during pregnancy and infancy SHS. However, there was an indication of an inverse association between allergic sensitization to outdoor allergens in children exposed to both maternal smoking during pregnancy and SHS in infancy, with a statistically significant reduced risk of outdoor allergen sensitization at 16 years of age (OR 0.68, 95% CI 0.47–0.96).

Table 1

Associations between maternal smoking during pregnancy (in utero) and parental smoking in infancy (SHS) in relation to sensitization at 4, 8, and 16 years of age among children in the BAMSE birth cohort (n = 3316)f

	N	Any allergena , e		Indoor allergensb , e		Outdoor allergensc , e		Food allergensd , e
		n	OR (95% CI)	n	OR (95% CI)	n	OR (95% CI)	n	OR (95% CI)
4 Years
No exposure in utero or SHS	1962	457	Referent	155	Referent	216	Referent	289	Referent
Exposure in utero but not to SHS	102	25	0.98 (0.61–1.58)	10	1.11 (0.54–2.25)	11	0.84 (0.43–1.64)	18	1.20 (0.71–2.03)
Exposure to SHS but not in utero	309	81	1.16 (0.88–1.54)	34	1.50 (1.00–2.24)	35	1.08 (0.74–1.59)	64	1.47 (1.08–2.00)
Exposure to in utero and SHS	219	61	1.16 (0.84–1.62)	25	1.41 (0.87–2.28)	30	1.14 (0.73–1.77)	40	1.17 (0.80–1.72)
8 Years
No exposure in utero or SHS	1845	642	Referent	328	Referent	397	Referent	353	Referent
Exposure in utero but not to SHS	91	33	1.03 (0.66–1.62)	19	1.14 (0.66–1.95)	22	1.07 (0.64–1.78)	19	1.12 (0.67–1.89)
Exposure to SHS but not in utero	305	104	0.97 (0.74–1.26)	51	0.93 (0.67–1.31)	58	0.88 (0.64–1.21)	75	1.34 (0.99–1.79)
Exposure to in utero and SHS	193	66	0.99 (0.72–1.38)	29	0.84 (0.55–1.29)	36	0.83 (0.56–1.23)	41	1.13 (0.78–1.65)
16 Years
No exposure in utero or SHS	1937	891	Referent	586	Referent	662	Referent	242	Referent
Exposure in utero but not to SHS	89	40	0.94 (0.60–1.46)	22	0.73 (0.44–1.22)	30	0.95 (0.60–1.52)	11	0.90 (0.46–1.77)
Exposure to SHS but not in utero	310	144	1.04 (0.81–1.33)	93	1.03 (0.79–1.35)	99	0.94 (0.72–1.22)	50	1.32 (0.93–1.86)
Exposure to in utero and SHS	195	88	0.99 (0.73–1.35)	54	0.90 (0.63–1.27)	49	0.68 (0.47–0.96)	28	1.19 (0.77–1.86)

Any positive inhalant and/or any food allergen.

Sensitization to cat, dog, horse, or mite.

Sensitization to timothy, birch, mugwort, or mold.

Sensitization to cow's milk, hen's egg, soy bean, peanut, cod fish, or wheat.

Adjusted for heredity and socioeconomic status.

Odds ratio (OR) and 95% confidence intervals (CI) obtained from multinomial logistic regression.

When assessing the longitudinal effects of SHS exposure, maternal smoking during pregnancy and parental smoking in infancy were analyzed separately to increase power and robustness. There were no overall increased risks of any, indoor, outdoor, or food allergen sensitization up to age 16 years among children exposed to maternal smoking during pregnancy (data not shown). In contrast, there was a suggestion of an overall increased risk of food sensitization up to age 16 years among children exposed to parental smoking in infancy (OR 1.24, 95% CI 0.98–1.56) (Fig. 2). Age‐specific associations were suggested for food allergen sensitization at all ages. These associations persisted after further adjustment for maternal smoking during pregnancy, although failing to reach statistical significance (OR 1.25, 95% CI 0.98–1.60 for the overall association). Furthermore, we observed an increased risk of any inhalant or food allergen as well as indoor inhalant allergen sensitization at age 4 years, but not at older ages or overall (Fig. 2).

Figure 2

Association between exposure to SHS during infancy and the risk of allergic sensitization during the first 16 years of life among children in the BAMSE birth cohort (n = 3316). **Any inhalant and/or any food allergen positive. †Sensitization to cat, dog, horse, or mite. ‡Sensitization to timothy, birch, mugwort, or mold. §Sensitization to cow's milk, hen's egg, soy bean, peanut, cod fish, or wheat. Adjusted for heredity, socioeconomic status, and smoking after infancy.

Stratified analyses showed no significant differences in SHS in infancy and food sensitization among those with and without parental allergic disease (OR 1.36, 95% CI 0.90–2.06 and OR 1.17, 95% CI 0.88–1.57, respectively) (Table S5). Moreover, we found no statistically significant interaction between SHS in infancy and parental allergic disease (P = 0.94). Likewise, these analyses were stratified by maternal smoking during pregnancy, and we observed excess risk in children only exposed in infancy (OR 1.33, 95% CI 1.01–1.75) compared to those exposed both in infancy and to maternal smoking during pregnancy (OR 0.97, 95% CI 0.57–1.64) (Table S5). No evidence of interaction was observed between maternal smoking during pregnancy and SHS in infancy (P = 0.30).

In analyses of single allergens, SHS exposure in infancy was associated with a statistically significant increased risk of sensitization to cow's milk (OR 1.32, 95% CI 1.02–1.70), and positive associations were also suggested for all other food allergens with the exception of cod fish (Table 2). An increased risk of mold sensitization in relation to infancy SHS exposure was also observed (OR 1.81, 95% CI 1.10–2.99).

Table 2

Associations between parental smoking in infancy (SHS) and the risk of sensitization to individual allergens up to 16 years of age among children in the BAMSE birth cohort (n = 3316)c

Allergena	Parental smoking in infancy
	OR (95% CI)b	P‐value
Birch
No SHS	Referent
Yes SHS	0.93 (0.74–1.16)	0.52
Timothy
No SHS	Referent
Yes SHS	0.81 (0.65–1.01)	0.06
Mugwort
No SHS	Referent
Yes SHS	0.71 (0.51–0.97)	0.03
Cat
No SHS	Referent
Yes SHS	1.12 (0.88–1.42)	0.35
Dog
No SHS	Referent
Yes SHS	0.98 (0.77–1.23)	0.84
Horse
No SHS	Referent
Yes SHS	1.31 (0.98–1.76)	0.07
Dust Mite
No SHS	Referent
Yes SHS	0.92 (0.68–1.23)	0.56
Mold
No SHS	Referent
Yes SHS	1.81 (1.10–2.99)	0.02
Cow's Milk
No SHS	Referent
Yes SHS	1.32 (1.02–1.70)	0.03
Peanut
No SHS	Referent
Yes SHS	1.26 (0.93–1.71)	0.14
Hen's Egg
No SHS	Referent
Yes SHS	1.20 (0.86–1.67)	0.29
Wheat
No SHS	Referent
Yes SHS	1.20 (0.87–1.67)	0.27
Soy bean
No SHS	Referent
Yes SHS	1.20 (0.83–1.73)	0.33
Fish
No SHS	Referent
Yes SHS	1.01 (0.38–2.68)	0.98

Sensitization defined as IgE ≥ 0.35 kUA/l.

Adjusted for heredity and socioeconomic status.

Analyses done with generalized estimating equations GEE).

When we explored the dose–response association between SHS exposure in infancy and the development of sensitization, we observed increased risks of any (OR 1.36, 95% CI 1.06–1.74), indoor (OR 1.42, 95% CI 1.05–1.91), and food allergen (OR 1.61, 95% CI 1.20–2.16) sensitization among children whose fathers smoked ≥10 cigarettes per day in infancy (Table 3). A significant trend of greater number of cigarettes was observed for any and food sensitization, P trend 0.04 and 0.003, respectively. Comparable associations were observed among children exposed to ≥10 cigarettes per day by both parents with any (OR 1.71, 95% CI 1.00–2.95) and food (OR 1.70, 95% CI 0.90–3.22) allergen sensitization.

Table 3

Associations between maternal smoking during pregnancy and parental smoking in infancy in relation to smoking intensity and overall risk of sensitization up to 16 years of age among children in the BAMSE birth cohort (n = 3316)a

	Maternal smoking during pregnancy	Parental smoking in infancy
	Motherb	Motherc	Fatherc	Both Parentsc
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Any allergend
No smoking	Referent	Referent	Referent	Referent
Total of 1–9 cigarettes/day	1.04 (0.76–1.42)	1.43 (0.98–2.06)	0.87 (0.65–1.16)	0.96 (0.42–2.18)
Total of ≥10 cigarettes/day	1.02 (0.76–1.36)	1.21 (0.82–1.75)	1.36 (1.06–1.74)	1.71 (1.00–2.95)
P trend	0.86	0.15	0.04	0.07
Indoor allergense
No smoking	Referent	Referent	Referent	Referent
Total of 1–9 cigarettes/day	0.76 (0.51–1.16)	1.18 (0.74–1.87)	0.69 (0.47–1.01)	0.58 (0.17–1.96)
Total of ≥10 cigarettes/day	1.09 (0.76–1.55)	1.13 (0.71–1.80)	1.42 (1.05–1.91)	1.30 (0.67–2.53)
P trend	0.96	0.49	0.10	0.63
Outdoor allergensf
No smoking	Referent	Referent	Referent	Referent
Total of 1–9 cigarettes/day	0.88 (0.60–1.29)	1.13 (0.73–1.77)	0.77 (0.54–1.09)	0.96 (0.36–2.59)
Total of ≥10 cigarettes/day	1.00 (0.71–1.42)	0.99 (0.63–1.56)	1.04 (0.77–1.39)	1.24 (0.65–2.38)
P trend	0.86	0.87	0.87	0.55
Food allergensg
No smoking	Referent	Referent	Referent	Referent
Total of 1–9 cigarettes/day	0.90 (0.61–1.33)	1.09 (0.68–1.74)	1.01 (0.71–1.45)	0.64 (0.19–2.17)
Total of ≥10 cigarettes/day	0.93 (0.66–1.33)	1.16 (0.73–1.84)	1.61 (1.20–2.16)	1.70 (0.90–3.22)
P trend	0.63	0.50	0.003	0.18

Odds ratio (OR) and 95% confidence intervals (CI) obtained from generalized estimating equations (GEE).

Adjusted for socioeconomic status, heredity, and postnatal smoking.

Adjusted for socioeconomic status, heredity, and smoking after infancy.

Any inhalant and/or any food allergen positive.

Sensitization to cat, dog, horse, or mite.

Sensitization to timothy, birch, mugwort, or mold.

Sensitization to cow's milk, hen's egg, soy bean, peanut, cod fish, or wheat.

Associations between SHS exposure in infancy in relation to asthma, rhinitis, and eczema in combination with sensitization to any common allergens are presented in Fig. 3, A. No significant associations were seen for asthma with or without sensitization. In contrast, a positive association was observed for rhinitis, but only among nonsensitized children (OR 1.53, 95% CI 1.01–2.31). We also observed elevated overall (OR 1.62, 95% CI 1.20–2.18) and age‐specific excess risks of eczema with any sensitization in relation to SHS exposure in infancy. The same analyses were performed among those children sensitized against food allergens (Fig. 3, B), providing similar results.

Figure 3

Association between parental smoking during infancy and the risk of any or food sensitization and allergic disease among children in the BAMSE birth cohort (n = 3316). **Any inhalant and/or any food allergen positive. §Sensitization to cow's milk, hen's egg, soy bean, peanut, cod fish, or wheat. Adjusted for heredity, socioeconomic status, and smoking after infancy.

In sensitivity analyses further adjustment for the children's own smoking at age 16 years had no major influence on the observed odds ratios between infancy SHS exposure and food sensitization (OR 1.28, 95% CI 1.00–1.64). Analyses were also performed with a higher antibody cut off of 3.5 kUA/l, and this did not significantly alter our findings (data not shown). Furthermore, analyses were conducted for inhalant, food, and single allergens using a reference group comprised of subjects without sensitization to any allergen, but showed no considerable differences.

Discussion

In this population‐based birth cohort followed up for 16 years, children exposed to parental smoking in infancy without prior exposure in utero had an increased risk for food allergen sensitization at age 4 years, with comparable associations at ages 8 and 16 years. Additionally, increased risk of indoor allergen sensitization at 4 years was observed among these children, but was no longer apparent at ages 8 or 16 years. A tendency of an overall association between SHS exposure in infancy and allergic sensitization to food allergens up to adolescence was suggested in longitudinal analyses. Moreover, there was a dose‐dependent association with an increasing number of cigarettes smoked by fathers in infancy. In contrast, there were no clear associations between maternal smoking during pregnancy and sensitization to inhalant or food allergens in the offspring. When we combined symptoms of allergic disease with sensitization status, we found that exposure to SHS in infancy was associated with excess risk of eczema with sensitization, but not eczema without sensitization.

Our finding that exposure to SHS in infancy is associated with an increased risk of sensitization to food allergens is in agreement with earlier studies in children between the ages of three and 4 years 7, 8. To our knowledge, our study is the first to indicate that the increased risk of food sensitization persists into adolescence. The German birth cohort, Multicenter Allergy Study (MAS), followed up to 10 years of age reported an increased risk of any sensitization among children exposed to regular maternal smoking, but only in children with a genetic predisposition 1. We observed similar risks of food sensitization irrespective of parental history of allergy.

A recent systematic review and meta‐analysis reported that exposure to postnatal parental smoking was associated with significantly higher total IgE concentrations, IgE antibodies to common allergens, and positive skin prick tests in preschool and school age children 16. However, the authors were unable to disentangle food sensitization from inhalant allergen sensitization. A study from the MAS birth cohort did differentiate between specific types of allergens as well as the effects of pre‐ vs postnatal parental smoking 8. They reported that preschool age children unexposed during pregnancy but exposed in early life had comparatively increased risks of food allergen sensitization as those children exposed both pre‐ and postnatally 8. In contrast, an earlier systematic review, as well as a UK birth cohort, found no associations between SHS exposure before or after birth and sensitization 17, 18. A lack of association between maternal smoking during pregnancy and sensitization in children at age 14 years was also shown in a recent report from an Australian birth cohort 19. In the present study, we observed no association with maternal smoking during pregnancy. In addition, we observed an attenuation of the risk in children exposed to both maternal smoking during pregnancy and parental SHS in infancy. One possible explanation for this is that exposure during pregnancy could preclude sensitization through an immunosuppressive effect 20, 21, 22, whereas exposure following birth confers risk of allergic sensitization 5. This could partly explain the divided literature on SHS exposure and sensitization, as some studies did not differentiate between pre‐ and postnatal SHS exposure.

We also observed an increased risk of mold sensitization in relation to infancy SHS exposure; however this finding should be interpreted with caution, as there were only 48 children sensitized against mold at age 16 years.

The presence of a dose–response association between SHS in infancy in our study and sensitization further reinforces our findings. Although the dose–response association was mainly observed in fathers, this may be because fathers were more likely than mothers to smoke and did so more heavily. Moreover, we observed larger risks of sensitization when both heavy smoking mothers and fathers were combined.

Although sensitization and allergic disease are interrelated, it is important to keep in mind that sensitization is not synonymous with allergic disease. Instead, the combination of symptoms and serologic measures of allergen‐specific IgE antibodies provided us with a more accurate picture of how SHS exposure affects allergic disease. In our study, we observed comparable associations between SHS exposure in infancy and asthma with and without concurrent sensitization. In contrast, an earlier study indicated a stronger association for asthma without sensitization 23. Focusing on any sensitization, SHS exposure in infancy was associated with an increased risk of rhinitis without sensitization in our study. Consistent with earlier studies in preschool age children, our study showed that SHS exposure in infancy increased the risk of eczema in combination with sensitization 24, 25.

The potential mechanisms behind the increased risk for sensitization in children of smoking parents are not fully understood. The ‘mucosal concept of atopy’ proposes that sensitization occurs mainly in the mucosal surfaces of the airways and that mucosal insults and inflammation facilitate antigen penetration and thus subsequent sensitization 26. Although it may seem puzzling that SHS should increase the risk of sensitization to food allergens, food allergens are frequently found in dust and are likely to be inhaled by infants. Infants often inhale or aspirate while eating and expose both the pharynx and pharyngeal lymphoid tissues to food allergens (as well as to tobacco smoke) 7. Furthermore, nicotine is known to have immunomodulatory effects and could thereby influence sensitization, although the exact mechanism is unknown 22. Alternatively, the dual‐allergen‐exposure hypothesis suggests that exposure through the skin leads to sensitization, whereas consumption of allergens leads to oral tolerance 27. As tobacco smoke has been shown to negatively affect skin barrier function and the ingress of allergens, food allergen sensitization is possibly influenced by the effect of SHS exposure on the skin 28, 29. Akin to our findings, studies exploring the risks of ambient traffic‐related air pollution – which may act through the same biological mechanisms as SHS – describe an increased risk of sensitization against food allergens associated with exposure in infancy 14, 30.

Several strengths of this study include its prospective and population‐based design, large size, high response rates, the use of validated questions based on the International Study of Asthma and Allergies in Childhood (ISAAC) 31, and questions used in other birth cohort studies 32. The collection of blood samples at three time points coupled with repeated questionnaires assessing SHS exposure enabled us to perform longitudinal analyses. Moreover, using Phadiatop® and fx5® we were able to detect greater than 90% of all sensitization in our study population. Information on maternal smoking during pregnancy and parental smoking in infancy were collected shortly after birth, thus reducing the likelihood of recall bias regarding smoking habits. Due to the large number of blood samples collected, we were able to differentiate between sensitization of specific inhalant and food allergens. The overall prevalence of sensitization increased with age and recent work from this same cohort indicates that the increase in sensitization is a dynamic process 33.

Limitations such as misclassification of exposure are often present in observational studies and should be acknowledged. SHS exposure was determined using questionnaires, so misclassification to some degree cannot be excluded. The prevalence of smoking may be underestimated due to the social stigmatization surrounding smoking, particularly in the context of children and health 34, and would lead to an attenuation of the true effect comparing dichotomous exposure categories. However, it is important to keep in mind that exposure information was assessed shortly following birth, thus parental responses would not be biased based on children's health or symptoms. Another potential limitation is the lack of information on paternal smoking during pregnancy assessed at baseline. Maternal passive smoking has been shown to increase the risk of allergic disease in children 35, 36, and could have an effect on IgE sensitization. Nevertheless, we saw no indication of an association between maternal smoking during pregnancy and sensitization, so it is unlikely that maternal passive smoking during pregnancy would significantly alter our findings.

In conclusion, our findings suggest that exposure to parental smoking in early infancy increases the risk of sensitization to food allergens up to 16 years. Exposure to SHS in infancy was also associated with an excess risk of eczema with sensitization. Exposure to maternal smoking during pregnancy was unrelated to sensitization up to 16 years of age.

Funding source

The Swedish Research Council, The Swedish Heart and Lung Foundation, The Swedish Research Council for Working Life and Social Welfare, the Swedish Asthma and Allergy Association Research Foundation, The Swedish Research Council Formas, Stockholm County Council, and the European Commission's Seventh Framework 29 Program MeDALL under grant agreement No. 261357. Thermo Fisher Scientific kindly provided the reagents for the study. None of the funding sources had a role in the study design, conduct, analysis, or reporting.

Conflict of interest

The authors have no potential conflicts of interest to disclose.

Author contributions

MW and GP initiated the BAMSE cohort. JDT, OG, GP, ÅN, MvH, MW, IK, EM, and AB participated in the design and planning of the current study. JDT performed the data analyses and drafted the initial manuscript. MvH supervised the analysis of blood samples. All authors participated in the interpretation of the findings, provided critical review, and approved the final manuscript as submitted.

Figure S1. Prevalence and proportional relationship of maternal smoking during pregnancy and parental smoking in infancy (n = 3290).

Figure S2. Distribution of food sensitization at 4, 8, and 16 years of age (n = 1684).

Table S1. Distribution of selected exposure characteristics among all children in the BAMSE cohort (n = 4089) and children who gave blood at ages 4, 8 and 16 years and those children included in the current analyses (n = 3316).

Table S2. Distribution of selected exposure characteristics in relation to parental smoking during infancy among children in the BAMSE birth cohort (n = 3316).

Table S3. Prevalence of sensitization at 4, 8, and 16 years of age among children in the BAMSE birth cohort (n = 3316).

Table S4. Associations between maternal smoking during pregnancy (in utero) and parental smoking in infancy (SHS) in relation to transient* and persistent† food allergen sensitization at 4, 8, and 16 years of age among children in the BAMSE birth cohort (n = 3316)**.

Table S5. Risk of food sensitization up to 16 years in relation to parental smoking in infancy stratified by parental allergic disease and maternal smoking during pregnancy among children in the BAMSE birth cohort (n = 3316)*.

Click here for additional data file.