In utero and childhood exposure to parental tobacco smoke, and allergies in schoolchildren

Population 

Out of 9000 children of fourth and fifth grade (9–11 years old) selected in six cities of metropolitan France (Bordeaux, Clermont-Ferrand, Créteil, Marseille, Reims, Strasbourg, respectively), 7798 participated into the survey that included an epidemiological questionnaire addressed to the parents on health, potential risk factors and management of allergic and respiratory diseases, a medical examination including skin prick test positivity to common allergens, skin examination for eczema and urticaria, and run test to assess EIA, in 2000. Each centre randomly selected 16 schools according to ISAAC protocol.

Authorization by the “National Commission of Informatics and Civil Liberties (CNIL)” and by the Ethic Committee was sought and obtained before conducting the survey. The parents of the children were informed by mail of the purposes and modalities of the survey, and their written assent was requested.

Questionnaire 

Parents filled out a self-administered standardized questionnaire at home. Main questions derived from the ISAAC questionnaire. Responses were used to assess the prevalence, severity and management of asthma and allied diseases, namely wheezing, allergic rhinitis, and eczema, respectively. Such questions had been previously validated and had been translated from English into French by a native French speaker, then back translated into English by a native English speaker. This process had been pursued until the initial English wording had been obtained. Final French wording was provided by consensus among the Allergy and Respiratory Medicine department heads of the participating centres.

The following health outcomes were considered in the analysis:

Environmental tobacco exposure was defined on the basis of parental active smoking habits as follows:

To take into account the cumulative exposure of the children to parental tobacco smoking, three subtypes of active tobacco smoking were defined, respectively, for the father, for the mother and for both:

The other factors taken into account as potential confounders or modifiers in the study of the relationship between smoking habits and allergic disorders were age, sex, geographic region, and family history of allergy, breast feeding, prematurity and birthweight obtained through the questionnaire and using the paediatric record.

Clinical tests 

Consent of children was obtained before clinical examination in the classrooms. Then they underwent physical examination in the class by a physician with height, weight, respiratory symptoms, skin prick test, respiratory function with measure of peak-flow (three measures were obtained). According to Williams’ protocol, atopic dermatitis was determined by questionnaire and physical examination. Fieldworkers were trained for respiratory function and physical examination. They run tests to assess EIA with a measure of peak flow after exercise. EIA was defined by diminution of 10% of PEF of the best values. An additional analysis was conducted for 15% for fall in PEF.

Skin prick tests (SPT) performed using Stallerpoints (Stallergens diagnostics). The allergens used for skin testing were cat, Dermatophagoïdes pteronyssinus, Dermatophagoïdes farinae (house dust mite), cat fur (Feld I), the outdoor mould genus Alternaria tenuis, blatta germanica, peanut, egg, fish, mixed grass and tree pollens; Histamine (1mg/ml histamine dihydrochloride) was used as a positive control and an uncoated lancet as the negative control. Tests were performed on the volar surface of the forearm using a standard template and the wheal size was recorded at 15min as the largest diameter and the diameter at 90° to its midpoint, each to the nearest whole millimetre. Mean wheal diameter was calculated as the average of the two diameters. Skin prick test positivity to a specific allergen was regarded as positive if the corresponding mean wheal diameter was greater than that of the negative control.

EIA challenge 

Baseline PEF was measured in all children who agreed. Before undergoing these lung function tests, subjects were advised to avoid β2-agonist or anticholinergic inhaler for 4h, and oral medications (β2-agonists, theophylline or anti-muscarinic agents) for 8h. When possible, children who reported a respiratory tract infection during the previous 3 weeks were rescheduled. Subjects were permitted five attempts to provide at least two technically acceptable manoeuvres. All those whose PEF attained at least 70% of the predicted value, were invited to undergo a exercise challenge test during 6min unless they reported that they had heart disease or epilepsy, or were taking a β2-blocker. Post-exercise PEF was recorded immediately after the challenge, 5, 10 and 15min later. Subsequently, if a fall in PEF of 10% was determined or if the child presented any respiratory symptom, the physician examined him at first, and second, a β2-agonist with inhalation chamber was administered in order to ensure the reversibility of the bronchospasm.

Statistical analysis 

Conventional methods were used for comparing percentages. The logistic regression model was used to determine the risk (odds ratio), crude and adjusted, respectively, between asthma and each allergic disorder and parental smoking habit, after adjustment for other potential risk factors. The population taken into account for the present analysis is represented by the overall group of 9–11 years old surveyed in all centres. Based on previous knowledge, we also studied gender, parental atopy, siblings, the length of breast feeding, day care outside the home, pet's ownership and SPT positivity as potential modifiers both by adding exposure-modifier interaction terms in the logistic regression analysis and by stratifying the analysis of the data. Sex ratio was defined as, i.e. male/female ratio. The analyses were performed using STATA.16

Response rate 

The final sample included 7798 children who participated to the study and 7242 children with complete records (questionnaire, clinical examination, skin prick test, exercise challenge). The child's biological mother for 81% of questionnaires completed the questionnaire. Missing data were excluded from analysis, however for each variable (i.e. low birth weight, premature birth, and lifetime asthma) we compared non-respondents and respondents according to demographic data, smoking exposure, and atopic status. We did not find any significant difference in these comparisons. Children were excluded for the following reason such as missing information on exposure to parental smoking. Most of the children were white and most had well-educated, middle-class parents. The response rate varied from 72% in Reims to 92% in Clermont-Ferrand. The sex ratio was well balanced in all six centres. There was no significant difference in age in the centres. The children with missing information on ETS exposure for specific smoking questionnaire were compared with children with complete data. Children with missing information on exposure to parental smoking were not different from the other children with respect to atopic disease in father and siblings, parental age, sex, season of birth, number of siblings, day care attendance and pet ownership. However, children with missing information on exposure to parental smoking were more likely to be asthmatic (36.8% versus 11%, P=0.0001), to have two parents with low level of education and with ethnic difference, not to belong to the European community. In particular, children with missing information on exposure to parental smoking were more frequent in the West Indies group (48%) than in the European group (15%, P=0.001).

In Table 1, the general characteristics of the study population are summarized, 25% had no siblings and 95.4% had social security. One parent out of three had a high education. As expected, current prevalence of asthma and allergies were lower than lifetime prevalence. Half of the children were exposed to parental tobacco smoking at home at some times in their lives. Maternal smoking increased with children age whereas paternal smoking diminished.

Table 1. General characteristics of the study population.

		Child	Mother	Father
	Age	10±0.8	38.4±5.2	41.4±6.42
	Male (%)	49.2
	Weight (kg) (M±sd)	36.5±12.6
	Height (m)(M±sd)	1.41±0.09
	BMI (M±sd)	18.38±12
	PEF (ml/mn) (M±sd)	330.46±49.41
	Ethnic group (%)
	White		68.13	62.7
	West Indies		3	2.99
	South Europe		4.67	4.9
	North-Africa		11.52	12.6
	Black Africa		3.45	3.3
	Asia		3	3.2
	Other		6.2	7.2
	Parental education (%)
	Primary school		13.5	14.2
	Secondary school		45.7	42.9
	Vocational training certificate		3.3	3.9
	Master degree standard		37.5	38.9
	Siblings
	No	25
	1 sister or brother	34.6
	2 sisters or brothers	22.7
	3 sisters or brothers	8.5
	>4 sisters or brothers	6
	Insurance family social security (%)	96.8
	Complementary private insurance (%)	1.2
	Lifetime asthma, n (%)	695(9.9)
	Current asthma, n (%)	509(7.9)
	Lifetime wheezing, n (%)	1365(19.4)
	Current wheezing, n (%)	518(7.4)
	Lifetime hay fever, n (%)	866(12.9)
	Allergic rhinitis, n (%)	732(11.6)
	Lifetime eczema, n (%)	1715(25.22)
	Exercise-induced asthma, n (%)	610(8.3)
	Atopy, n (%)	1848(25)	1673(21.5)	1376(17.65)
	Breast feeding		3710(54.4)
	Birthweight <2500g	443(6.8)
	At least one sibling, n (%)	2705(34.6)
	Low birthweight birth (<37 weeks of gestation), n (%)	1879(27.9)
	Day care at least 6 months, out home, n (%)	1392(22.1)
	Exposure to pets during the first year of life, n (%)	504(6.9)

Asthma and allied diseases 

Eczema was the most frequent disorder among the studied children followed up by allergic rhinitis and asthma (Table 1). A history of asthma at some point in life and of current wheezing (current asthma) was reported by 7.9%. There were significantly more boys who had lifetime asthma and current wheezing and asthma, nasal allergic symptoms and lifetime and current eczema. Asthma and allied diseases were more frequent in the two southern French centres than in the two northern ones, with maximum prevalence in Marseille. 8.3% of children had an EIA according to the challenge performed in the school. 25% had skin prick test positivity. 21.5% of children had a mother with personal atopy (i.e. asthma or hay fever or eczema). 54.4% of mothers reported breast feeding for their child, however 46.7 of them (i.e. n=1286) reported breast feeding for at least 3 months and 47.4% inferior to 1 month.

Exposure to parental smoking 

About 32% of children were exposed to parental smoking during pregnancy. 43% were exposed to parental smoking during the first year of life, and 38.6% were currently exposed during the study. The prevalence of exposure to parental smoking rose during life with higher prevalence when children are older. The number of cigarettes smoked per day and declared by the parents, increased also with age of the children (Table 2), and was lower for the mother than for the father. Overall, father smokers consumed five cigarettes a day compared to 1.9 for mother smokers during pregnancy. Correlations between maternal and paternal smoking compared with never-smokers are presented in Table 3. About 786 children were exposed to other sources of smoking according to parents. The majority of exposed children were exposed to maternal smoking too.

Table 2. Prevalence of parental smoking and number of cigarettes smoked.

		N (%)	Number of cigarettes per day M(sd)
	In utero (responders)
	Mother (5532)	1193(21.6)	1.9(5.4)
	Father (4588)	1721(37.51)	5(8.9)
	Other (2296)	111(4.6)	0.58(3.59)
	Mother and father (4442)	540(12)	6.2(10.9)
	First year of life
	Mother (5442)	1619(29.75)	3.5(7.44)
	Father (4957)	2028(40.91)	5.54(9.77)
	Other (2511)	178(7)	0.91(5.7)
	Mother and father (4785)	977(20)	8.4(14)
	Current
	Mother (6042)	2060(34.1)	3.9(7.5)
	Father (5415)	1958(36.16)	4.6(9.2)
	Other (2822)	345(12.2)	1.34(5.4)
	Mother and father (5200)	1017(19.5)	7.5(13.1)

Table 3. Correlations between paternal and maternal smoking, OR Ic95%.

		Never smoked	Maternal smoking only	Paternal smoking only	Both maternal and paternal smoking
	In utero	1	2.4[2.1–2.7]	1.38[1.3–1.43]	5.8[4.9–6.9]
	First year	1	3.08[2.7–3.4]	1.8[1.7–1.9]	8.9[7.7–10.4]
	Current exposure	1	2.4[2.2–5.8]	1.9[1.8–2.1]	7.7[6.7–8.8]

Boys were more exposed to current paternal smoking than girls (sex ratio: 1.14[1.02–1.27], P=0.01), and girls were more exposed to maternal smoking during the first year than boys (sex ratio: 0.88[0.79–0.99], P=0.04). Taking into account exposure to both maternal and paternal smoking during life, boys were significantly more exposed than girls (sex ratio: 1.35[1.06–1.73], P=0.01). There was a significant association between passive exposure and a birthweight <2500g (odds ratio (OR): 1.79[1.39–2.5]) with exposure-response pattern with the number of cigarettes smoked per day (>20cig/d; OR: 2.91[1.11–7.62]). There was also a significant association between passive exposure during pregnancy and prematurity (OR: 1.2[1.04–1.38]). In utero exposure was significantly associated with low birth weight (9.4% vs. 5.4%; P=0.001), however low birth weight was not associated with lifetime wheeze, but was associated with asthma (12.4% vs. 9.3%: P=0.001).

We adjusted the model for low birth weight, in addition with others potential confounders, the association remains significant between smoking exposure and wheeze. Children with low birth weight were mostly premature (73.4%); however, not all premature children were of low birth weight.

Asthma and allergies related to passive exposure 

ETS during pregnancy was associated with both lifetime wheezing and asthma (Table 4). Smoking, both paternal and maternal, during pregnancy and during the first year of life of the child was associated with lifetime asthma. Maternal ETS was associated with both lifetime asthma and wheezing but there was no dose-effect cumulative exposure. Relations were stronger in girls. No relation was found between current wheezing and ETS exposure (Table 4), the same result was observed for allergic rhinitis. Surprisingly, an inverse relation was found between on one hand, current exposure and lifetime hay fever, and on the other hand, current exposure and lifetime eczema. EIA of children was associated with current paternal smoking exposure. No relationship was found between SPT positivity of the children and parental tobacco exposure.

Table 4. Odds ratios (and 95% confidence intervals) between asthma and allergic disorders and exposure to parental tobacco smoke, according to prior history of exposure.

*P<0.05, ***P<0.001. Adjusted on gender, siblings, exposure to pets during the first year of life, ethnic group, parental education, day care, breast feeding, geographic center, parental atopy, prematurity. *P<0.005 ref=1 for no-exposed and no asthma or allergies.

Taking into account the number of cigarettes smoked by parents per day (Table 5), smoking at least 10 and more than 20 cigarettes per day during pregnancy was significantly associated with lifetime wheezing showing the existence of a trend. No relation was found between exposure to parental smoking, current exposure particularly, and current wheezing. Similarly, there was no association for lifetime asthma. EIA at the age of 10 was significantly associated with parental tobacco smoking more than 20cig/d during the first year of life. We did not observe an exposure-response pattern for any outcome variables regarding the number of cigarettes smoked per day.

Table 5. Odds-ratio between wheezing and asthma and the number of cigarettes smoked per a day.

		Lifetime wheezing (4386)	Wheezing during the past year (1854)	Lifetime asthma (5136)	Exercise-induced asthma (5124)
	Cigarettes smoked per day (mother and father) during pregnancy
	0(2610)	1	1	1	1
	1–9(939)	1.28[1.07–1.53]***	0.96[0.7–1.31]	1.08[0.84–1.38]	1.09[0.83–1.44]
	10–19(544)	1.08[0.86–1.36]	0.84[0.56–1.28]	0.83[0.59–1.16]	0.95[0.66–1.36]
	>20 (349)	1.35[1.04–1.77]*	1.28[0.83–1.98]	1.15[0.8–1.65]	1.42[0.97–2.06]
	P for trend	0.02	0.52	0.45	P=0.30
	Cigarettes smoked per day (mother and father) during first year of life
	0(2583)	1	1	1	1
	1–9(960)	1.28[1.07–1.53]***	0.95[0.70–1.29]	1.15[0.9–1.47]	1.14[0.86–1.5]
	10–19(664)	1.05[0.85–1.3]	0.70[0.47–1.04]	0.82[0.6–1.12]	1.01[0.72–1.39]
	>20(578)	1.14[0.91–1.42]	1.22[0.85–1.73]	1.25[0.94–1.67]	1.41[1.04–1.93]*
	P for trend	0.05	0.14	0.10	0.16
	Cigarettes smoked Currently per day (mother and father)
	0(2914)	1	1	1	1
	1–9(1060)	1.07[0.9–1.28]	1.02[0.76–1.37]	1.15[0.91–1.45]	0.99[0.75–1.31]
	10–19(660)	1.03[0.83–1.28]	0.97[0.67–1.39]	0.97[0.72–1.3]	1.22[0.89–1.67]
	>20(566)	1.04[0.82–1.3]	0.89[0.61–1.30]	1.03[0.76–1.40]	1.30[0.94–1.8]
	P for trend	0.87	0.93	0.66	0.28

Adjusted on gender, siblings, exposure to pets during the first year of life, ethnic group, parental education, day care, breast feeding, geographic center, parental atopy, prematurity. The number given under each symptom indicates the number of subjects included in the analysis of that outcome variable, beside exposure, the number indicates the number of subjects with that exposure status. *P<0.05. ***P<0.001.

Modifier effects 

We tested whether the effect of in utero exposure to smoking on asthma and wheeze could be mediated through prematurity.

The results of this analysis are now presented in Table 6.

Table 6. Associations between in utero exposure and asthma and wheeze by prematurity status.

		Prematurity (n=1879)			Born at term (n=4866)
		Smoking exposure (%)	No-smoking exposure (%)	Adjusted OR Ic 95%	Smoking exposure (%)	No-smoking exposure (%)	Adjusted OR Ic 95%
	Wheeze	29.3	25	1.24 [0.95–1.62]	23.8	18	1.23[1.01–1.49]
	Asthma	15.7	11.4	1.44 [1.03–2.01]	9.9	9.1	1.01[0.84–1.4]

Adjusted on gender, siblings, exposure to pets during the first year of life, ethnic group, parental education, day care, breast feeding, geographic center, parental atopy.

Relations between ETS exposure during pregnancy and lifetime wheezing were stronger in children with atopic parents (OR: 1.8[1.33–2.4]). Relations between ETS exposure and lifetime wheezing were significantly stronger in children with negative SPT to pollens or Alternaria (OR: 1.4[1.13–1.73]). Unlike with indoor allergens, the relations between ETS exposure and lifetime wheezing were stronger in children sensitized to house dust mites and cat (OR: 1.71[1.18–2.5]). Breast feeding did not modify the relations between exposure and the considered health outcomes. Owing pets during the first year of life had a modifier effect on the relation between smoking exposure and lifetime wheezing, with a stronger effect in children without pets (OR: 1.35[1.08–1.68]). In the same way, day care during the six first months of life had a stronger effect on the relations between exposure and considered outcomes (OR: 1.4[1.07–1.84]).

Children whose mothers smoked during the first year of life but did not smoke during pregnancy, had a reduced risk to have asthma (OR: 1.1[0.82–1.48]) and wheeze (OR: 1.08[0.86–1.35]) compared with others children, after adjustment on potential confounders.