Effect of Secondary Cigarette Smoke from Household Members on Breastfeeding Duration: A Prospective Cohort Study

Abstract

Objective:

Maternal smoking and exposure to tobacco smoke is one modifiable risk factor that affects breastfeeding initiation and duration. We prospectively examine the effects of maternal and paternal smoking and other family members' smoking status on the duration of breastfeeding.

Materials and Methods:

A cohort of 1,277 mother–infant pairs admitted to the postpartum units of four major public hospitals in Hong Kong. Demographic data, maternal, paternal and household smoking habits, and other potential confounding variables were collected via self-reported questionnaires during the postnatal hospitalization. Breastfeeding status after hospital discharge was assessed through telephone follow-up up to 12 months postpartum, or until participants were no longer breastfeeding. If the participant had weaned during that follow-up interval, she was asked to report the total duration (in weeks) of both any and exclusive breastfeeding.

Results:

A total of 1,240 (97%) mother–infant pairs completed followed up, 2.5% were smokers, 29.2% of their partners smoked, and 11.3% had another smoker living in their home. Maternal and other family members' smoking predicted breastfeeding cessation. When compared with mothers in nonsmoking family, those exposed to two or more family members who smoked had approximately a 30% increased risk of breastfeeding cessation (adjusted hazard ratio [aHR] = 1.31; 95% CI 1.01–1.68).

Conclusion:

Mothers who were exposed to two or more smokers in the household had a significantly shorter duration of any breastfeeding at 12-month follow-up.

Introduction

Breastfeeding provides the best nutrition for optimal growth and development. The World Health Organization recommends 6 months of exclusive breastfeeding with continued breastfeeding for up to 2 years and beyond.¹ Breastfeeding confers many benefits to infants, such as reduced risk of respiratory diseases, otitis media, anemia, digestive system diseases, secondary infections in preterm babies, and sudden infant death syndrome.^2,3 Given these benefits, the proportion of new mothers in developed countries choosing breastfeeding has increased substantially over the past several decades. However, worldwide, only 39% of infants are exclusively breastfed for at least 6 months and in East Asia that number drops to 31%.⁴ Therefore, it is important to identify modifiable determinants of breastfeeding exclusivity and duration in various populations to understand barriers to breastfeeding practices.

Maternal smoking and exposure to tobacco smoke is one modifiable factor that affects breastfeeding initiation and duration.⁵ Hong Kong is a crowded city with high population density. Living with extended family members is common. Although over 86% of new mothers initiate breastfeeding,⁶ exclusive breastfeeding rates at 6 months are only 0.9%.⁶ Because of the crowded living conditions, many pregnant women and young children are exposed to environmental tobacco smoke from household and family members. Recent research showed that while only 3.4% of Hong Kong pregnant women smoke, 33% of their partners' smoke.⁷ Furthermore, women who smoke during pregnancy or who have smoking partners are significantly less likely to initiate breastfeeding.^7,8 There have been limited studies in Hong Kong examining the effects of maternal, paternal, and household smoking patterns on breastfeeding duration.^7–9 Leung et al.^8,9 failed to find a consistent relationship between smoking patterns and breastfeeding duration. An explanation to this was that smoking habits may reflect a person's attitude toward health and less enthusiastic about breastfeeding than their nonsmoking counterparts. Studies in other populations have shown that female smokers are less likely to breastfeed than nonsmokers, and women who smoke and breastfeed are more likely to breastfeed for a shorter duration.^10–12 Women who stop smoking during pregnancy may stop breastfeeding early to resume smoking. Researchers found that 50% of women resumed smoking by 4 months postpartum, and most mothers who resumed smoking had a smoking partner and were more likely to stop breastfeeding before 6 weeks postpartum.^13,14

Previous Hong Kong studies examining the effect of smoking on breastfeeding duration were conducted more than 20 years ago.^8,9 With the implementation of strong tobacco control measures in Hong Kong since 2007,¹⁵ population-level smoking rates have decreased substantially. In addition, breastfeeding supportive policies and promotional efforts have improved breastfeeding initiation and duration.^16,17 Therefore, the aim of this study is to examine the effects of maternal, paternal, and household smoking on the duration of any and exclusive breastfeeding.

Materials and Methods

Data source and sample

This study was part of a large longitudinal cohort study assessing the effect of hospitals accepting free infant formula on breastfeeding duration and exclusivity in infants. The study methods are described in more detailed elsewhere.¹⁶ In brief, we recruited 1,287 mother–infant pairs from the postpartum obstetric units of four public hospitals in Hong Kong. Participants were recruited from October 2011 to July 2012. There are 8 public and 10 private hospitals that provide obstetric services in Hong Kong. In 2016, 67% of all Hong Kong mothers gave birth in public hospitals.¹⁷ The following criteria were used for the selection of study participants: (1) intention to breastfeed; (2) women with singleton pregnancies; (3) Cantonese speaking; (4) Hong Kong residents for more than 12 months; and (5) no serious medical or obstetrical complications. Nonbreastfeeding participants were not recruited. Participants were excluded from the study if the infant (1) was born before 37 weeks' gestation, (2) had an Apgar score of less than eight at five minutes, (3) had a birthweight of less than 2,500 g, (4) was born with any severe medical conditions or congenital malformations, (5) was placed in the special care nursery for more than 48 hours after birth, or (6) was placed in the intensive care nursery after birth. After hospital discharge, participants were followed-up by telephone for 12 months or until breastfeeding had stopped, whichever was earlier.

Basic sociodemographic and smoking status information were self-reported by participants during their postpartum hospital stay. Maternal and birth data were collected by a trained research nurse, also during the postpartum hospitalization. Data were collected on breastfeeding status in follow-up telephone interviews up to 12 months after birth. At each follow-up, the pattern of infant feeding was classified as either exclusive or partial breastfeeding, with clear explanations of each provided.^18,19 At the follow-up interview after breastfeeding had stopped, participants were asked to report the number of weeks of any and exclusive breastfeeding. No further data were collected after this point. Participants who were still breastfeeding at the 12-month follow-up did not complete the final breastfeeding data.

Study variables

The primary outcomes were the overall duration of any and exclusive breastfeeding, measured in weeks up to 12 months postpartum. The primary independent variables were the smoking status of the participant (maternal smoker), their partner (paternal smoker), and household members (household smoker). Participants were asked whether they smoked during pregnancy, whether their partner smoked, and whether any other family members living in the same household smoked. The smoking status of the participant, the partner, and household members was further categorized as the total number of smokers living in the household (0, 1, and ≥2). In addition, we measured well-known confounding variables, including maternal age, maternal education, household income, length of residence in Hong Kong, returning to work postpartum, and partner's infant feeding preferences.

Statistical analysis

Descriptive statistics and chi-square tests were used to describe and compare the characteristics of study participants by the three measured smoking variables (maternal, paternal, and household smoking). Kaplan–Meier survival curves and log-rank tests (trend) were performed to explore the association between the number of household smokers and total breastfeeding duration. We used unadjusted and adjusted Cox proportional hazards models to evaluate the extent to which smoking status influenced breastfeeding duration. In the multivariable analyses, we adjusted for the sociodemographic variables identified above that have been shown to affect breastfeeding duration in this and other populations.^20,21 For all analyses, participants with partial follow-up were censored at the time of the last follow-up. All data analysis was conducted using Stata version 14.2 statistical software (Stata Corp, College Station, TX,).²² A 0.05 nominal level of statistical significance was used throughout the analysis. Ethical approval for the study was obtained from the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster and all participating study sites. Informed written consent was obtained from all participants.

Results

Sample characteristics

A total of 1,287 mother–infant pairs were in the baseline sample. After recruitment, we excluded 10 participants because they met a study exclusion criteria (i.e., transferred to the neonatal intensive care unit), and 37 (2.9%) participants with no further contact after hospitalization. We included 1,240 mother–infant pairs in the final analysis. Of these, 1,201 participants had full follow-up until they stopped breastfeeding, and 39 (3.1%) had partial follow-up. Of the 1,240 participants, 2.5% were smokers, 29.2% of their partners smoked, and 11.3% had another smoker living in their home. In total, 34.8% (n = 432) of participants lived in a smoking household.

The characteristics of the study participants according to smoking status are shown in Table 1. There were significant differences in the age of participants, education level, family income, length of residence on Hong Kong, return to work postpartum, and partner's infant feeding preference. Participants who smoked were more likely to be younger, have a lower education level and have a lower family income, and less likely to be returning to work postpartum. Participants with smoking partners were also more likely to have similar characteristics and the partners were more likely to prefer infant formula or mixed feeding for the infant.

Table 1.

A Comparison of the Characteristics of Study Participants by Maternal, Paternal, and Household Member Smoking Status

	Maternal smoker			Paternal smoker			Any household smoker
Characteristic	No (N = 1209) N (%)	Yes (N = 31) N (%)	p	No (N = 878) N (%)	Yes (N = 362) N (%)	p	No (N = 808) N (%)	Yes (N = 432) N (%)	p
Maternal age (years)			<0.001			<0.001			<0.001
18–24	63 (5.2)	10 (32.3)		29 (3.3)	44 (12.2)		17 (2.1)	56 (13.0)
25–29	298 (24.7)	10 (32.3)		201 (22.9)	107 (29.6)		182 (22.5)	126 (29.2)
30–34	523 (43.3)	6 (19.4)		403 (45.9)	126 (34.8)		378 (46.8)	151 (35.0)
≥35	325 (26.9)	5 (16.1)		245 (27.9)	85 (23.5)		231 (28.6)	99 (22.9)
Maternal education			0.001			<0.001			<0.001
Compulsory secondary/below	29 (2.4)	2 (6.5)		13 (1.5)	18 (5.0)		10 (1.2)	21 (4.9)
Upper secondary	675 (55.8)	26 (83.9)		433 (49.3)	268 (74.0)		383 (47.4)	318 (73.6)
University degree or above	505 (41.8)	3 (9.7)		432 (49.2)	76 (21.0)		415 (51.4)	93 (21.5)
Monthly family income (HKD)^a			<0.001			<0.001			<0.001
<$15,000	197 (16.3)	15 (48.4)		121 (13.8)	91 (25.1)		100 (12.4)	112 (25.9)
$15,000–$29,999	427 (35.3)	16 (51.6)		279 (31.8)	164 (45.3)		253 (31.3)	190 (44.0)
≥$30,000	585 (48.4)	0 (0.0)		478 (54.4)	107 (29.6)		455 (56.3)	130 (30.1)
Length of time in Hong Kong			0.646			0.223			0.070
< 5 years	89 (7.4)	3 (9.7)		58 (6.6)	34 (9.4)		50 (6.2)	42 (9.7)
5–≥15 years	403 (33.3)	8 (25.8)		296 (33.7)	115 (31.8)		268 (33.2)	143 (33.1)
Since birth	717 (59.3)	20 (64.5)		524 (59.7)	213 (58.8)		490 (60.6)	247 (57.2)
Return to work postpartum			0.001			<0.001			<0.001
No	421 (34.8)	20 (64.5)		281 (32.0)	160 (44.2)		246 (30.5)	195 (45.1)
Yes	788 (65.2)	11 (35.5)		597 (68.0)	202 (55.8)		562 (69.6)	237 (54.9)
Previous breastfeeding experience			0.102			0.386			0.189
No	721 (59.6)	23 (74.2)		520 (59.2)	224 (61.9)		474 (58.7)	270 (62.5)
Yes	488 (40.4)	8 (25.8)		358 (40.8)	138 (38.1)		334 (41.3)	162 (37.5)
Husband's infant feeding preference			0.292			<0.001			0.007
Breastfeeding	519 (42.9)	9 (29.0)		404 (46.0)	124 (34.3)		370 (45.8)	158 (36.6)
Infant formula and mixed feeding	82 (6.8)	3 (9.7)		60 (6.8)	25 (6.9)		52 (6.4)	33 (7.6)
No preference	608 (50.3)	19 (61.3)		414 (47.2)	213 (58.8)		386 (47.8)	241 (55.8)

1 USD = 7.78 HKD.

HKD, Hong Kong Dollar.

The unadjusted and adjusted associations between smoking status and the duration of any and exclusive breastfeeding are presented in Table 2. In the adjusted model, when compared with participants who were nonsmokers and had nonsmoking partners, participants with a smoking partner had 23% increased risk of breastfeeding cessation (adjusted hazard ratio [aHR] 1.23, 95% CI 1.06–1.42) and participants who smoked and who had a smoking partner had a 64% increased risk of breastfeeding cessation (aHR 1.64, 95% CI 1.07–2.51). Maternal smoking alone did not significantly affect the duration of any breastfeeding and there was no significant association between smoking by any family member and the duration of exclusive breastfeeding.

Table 2.

Unadjusted and Adjusted Hazard Ratios for Any and Exclusive Breastfeeding Cessation by Smoking Status

		Any breastfeeding				Exclusive breastfeeding
		Unadjusted		Adjusted		Unadjusted		Adjusted
Smoking status	Prevalence (%)	HR (95% CI)	p	HR (95% CI)^a	p	HR (95% CI)	p	HR (95% CI)^a	p
Both nonsmokers	70.4	1.00		1.00		1.00		1.00
Maternal smoker	2.5	0.64 (0.21–2.00)	0.445	0.54 (0.17–1.68)	0.286	0.45 (0.17–1.20)	0.110	0.43 (0.16–1.17)	0.099
Paternal smoker	29.2	1.34 (1.17–1.54)	<0.001	1.23 (1.06–1.42)	0.007	1.05 (0.93–1.96)	0.417	1.06 (0.93–1.21)	0.389
Both smokers	2.1	1.99 (1.31–3.02)	0.001	1.64 (1.07–2.51)	0.024	1.11 (0.75–1.65)	0.589	1.08 (0.72–1.62)	0.686

Adjusted for maternal age, maternal education, household income, length of residence in Hong Kong, returning to work, and husband's infant feeding preference.

HR, hazard ratio.

The relationship between the number of smokers in the household and the duration of any breastfeeding is shown in Figure 1. When compared with participants living in nonsmoking households, the risk of breastfeeding cessation was progressively higher in participants with more smokers in the household. The unadjusted and adjusted associations between the number of household smokers and the duration of any and exclusive breastfeeding are presented in Table 3. In the adjusted analysis, the linear relationship between the number of household smokers and the risk of breastfeeding cessation remained. When compared with participants who lived in nonsmoking households, participants exposed to two or more household smokers had an approximately a 30% increased risk of weaning (aHR = 1.31; 95% CI 1.01–1.68). The number of household smokers did not significantly affect the duration of exclusive breastfeeding.

FIG. 1.

The relationship between the number of smokers in the household and the duration of any breastfeeding.

Table 3.

Unadjusted and Adjusted Hazard Ratios for Any and Exclusive Breastfeeding Cessation by the Total Number of Smokers in the Household

		Any breastfeeding				Exclusive breastfeeding
		Unadjusted		Adjusted		Unadjusted		Adjusted
Total no. of smokers	Prevalence (%)	HR (95% CI)	p	HR (95% CI)^a	p	HR (95% CI)	p	HR (95% CI)^a	p
None	65.1	1.00		1.00		1.00		1.00
1	27.2	1.28 (1.12–1.48)	<0.001	1.21 (1.05–1.41)	0.011	1.04 (0.91–1.18)	0.553	1.07 (0.93–1.22)	0.331
≥2	7.7	1.65 (1.31–2.08)	<0.001	1.31 (1.01–1.68)	0.038	1.11 (0.89–1.38)	0.346	1.08 (0.86–1.37)	0.505

Adjusted for maternal age, maternal education, household income, length of residence in Hong Kong, returning to work, and husband's infant feeding preference.

Discussion

Prevalence of maternal smoking (2.5%) is low in this study, which may explain the reduced power to detect any association between maternal smoking and breastfeeding duration. In contrast to maternal smoking, paternal smoking was negatively associated with rates of any breastfeeding. This is consistent with previous studies conducted in Mainland China,^8,23 although the prevalence of paternal smokers in Hong Kong (∼30%) is lower than that in China (∼65%), because male smoking is part of the Chinese culture. In this study, participants whose partners smoked were 20% more likely to stop breastfeeding compared with nonsmoking participants with nonsmoking partners. Similarly, if both participants and their partners smoked, they were 60% more likely to stop breastfeeding compared with nonsmoking participants with nonsmoking partners. This has substantial health and economic effects on the infants and families in the study. The higher breastfeeding cessation rate in paternal smoking families found in our study is similar to rates reported in previous studies on maternal smoking and breastfeeding.^8,23

Nicotine has the physiological effect of lowering prolactin levels,^24,25 thereby potentially reducing milk production. Nicotine also increases adrenaline levels, which may result in vasoconstriction and impaired milk ejection.²⁶ Consistent with previous studies,^8,9 we found that exposure to household smokers also had a substantial negative effect on breastfeeding practices. More than one-third of participants had partners or other household members who smoked. In addition, fathers who smoked were significantly more likely to prefer infant formula or mixed feeding when compared with nonsmoking partners. Studies also show that a partner's infant feeding preference for infant formula substantially reduces breastfeeding duration.^27,28 Therefore, smoking partners may have affected the decision to stop breastfeeding. Our study findings support this hypothesis and suggest that paternal and household smoking exposure is strongly associated with a shorter breastfeeding duration.^8,23

This study is one of the first to examine the effect of family members' smoking on the duration of breastfeeding in Hong Kong after substantial changes to tobacco control regulations in 2007. Ours was a multicenter study that recruited participants from four large publicly funded hospitals in Hong Kong. We conducted regular telephone follow-up of the participants to minimize recall bias of breastfeeding status. Study attrition was low, and we had follow-up data on 96.3% of the original cohort. Despite the large sample size, it was not population based. Thus, new mothers with more breastfeeding confidence may have been more likely to agree to participate in this study, and we do not have data on eligible mothers who chose not to participate. Furthermore, smoking status data and breastfeeding outcomes may be subject to recall bias as they were self-reported by participants and other objective measures were not used. However, other studies have shown that maternal reports of breastfeeding duration are accurate for many years after women have stopped breastfeeding,²⁹ and maternal and household smoking patterns reported here are also consistent with other studies in this population.⁷ We also did not assess direct exposure and proximity to tobacco smoke. Thus, having smokers in the home is a crude indicator of tobacco smoke exposure and may or may not accurately reflect exposure levels. We also did not assess other lifestyle or environmental factors that are often associated with smoking in males that may confound the results such as body weight, height, diet, and occupational exposures.³⁰

It is likely that a substantial proportion of participants whose partners and household members smoked were also exposed to environmental tobacco smoke during pregnancy and infants may have been exposed after birth. Environmental tobacco smoke exposure increases the frequency and severity of respiratory infections,^31,32 which in turn could disrupt breastfeeding. Furthermore, environmental tobacco smoke can be present long after visible smoke vanishes; often continuing for months after the burned tobacco is extinguished.³³ In recent years, public health professionals have started using the term “third hand smoke” to describe smoke residues that linger on surfaces, fabrics and in dust after smoking.³³ Involuntary tobacco smoke exposure arises from numerous sources of exposure poses higher health risks.³⁴ Antenatal care providers in Hong Kong need to increase efforts to offer smoking cessation programs to pregnant women who continue to smoke during pregnancy and to partners and family members who smoke to reduce environmental tobacco smoke exposure for the new mother and infant.

Conclusion

Findings from this study provide evidence that mothers who are exposed to a greater number of smokers in the household have a significantly higher risk of breastfeeding cessation. Practitioners can assess the smoking patterns of pregnant women and their family members to provide smoking cessation education and support this high-risk group. Further research is needed to dissect the mechanisms of smoking on breastfeeding practices.

Funding

This study was supported by the Health and Medical Research Fund (grant no. 05060721) of the Food and Health Bureau, Government of the Hong Kong Special Administration Region.

Ethics Approval

The study has ethics approval from the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster and all participating study sites, Hong Kong.

Authors' Contributions

All authors contributed to drafts and approved of the final version of the article. K.Y.W.L. wrote the article and undertook the data analysis. V.H.S.C. undertook data cleaning, telephone follow-up, and performed initial data analysis. M.P.W. revised the article and data analysis section. M.T. led the study design, obtained funding, and revised the article.

Patient Consent

Obtained.

Footnotes

Disclosure Statement

No competing financial interests exist.

References

World Health Organization. Global Strategy for Infant and Young Child Feeding. Geneva, World Health Organization, 2003.

Rollins

, Bhandari

, Hajeebhoy

, et al. Why invest, and what it will take to improve breastfeeding practices?. Lancet, 2016; 387:491–504.

Victora

, Bahl

, Barros

, et al. Breastfeeding in the 21st century: Epidemiology, mechanisms, and lifelong effect. Lancet, 2016; 387:475–490.

UNICEF. The State of the World's Children 2016. Geneva, UNICEF, 2016.

Karmaus

, Dobai

, Ogbuanu

, et al. Long-term effects of breastfeeding, maternal smoking during pregnancy, and recurrent lower respiratory tract infections on asthma in children. J Asthma, 2008; 45:688–695.

Department of Health. Breastfeeding Survey 2017. 2017. Available at: www.fhs.gov.hk/english/reports/files/BF_survey_2017.pdf (accessed January 29, 2018).

Lok

, Bai

, Tarrant

. Predictors of breastfeeding initiation in Hong Kong and Mainland China born mothers. BMC Pregnancy Childbirth, 2015; 15:286.

Leung

, Ho

, Lam

. Maternal, paternal and environmental tobacco smoking and breast feeding. Paediatr Perinat Epidemiol, 2002; 16:236–245.

Leung

, Lam

, Ho

. Breast-feeding and its relation to smoking and mode of delivery. Obstet Gynecol, 2002; 99:785–794.

10.

Joseph

, Emery

, Bogen

, et al. The influence of smoking on breast feeding among women who quit smoking during pregnancy. Nicotine Tob Res, 2017; 19:652–655.

11.

Donath

, Amir

, Team

. The relationship between maternal smoking and breastfeeding duration after adjustment for maternal infant feeding intention. Acta Paediatr, 2004; 93:1514–1518.

12.

Haslam

, Lawrence

, Haefeli

. Intention to breastfeed and other important health-related behaviour and beliefs during pregnancy. Fam Pract, 2003; 20:528–530.

13.

Lauria

, Lamberti

, Grandolfo

. Smoking behaviour before, during, and after pregnancy: The effect of breastfeeding. Sci World J, 2012; 2012:1–9.

14.

, Wen

, Rissel

, et al. Smoking status and factors associated with smoking of first-time mothers during pregnancy and postpartum: Findings from the healthy beginnings trial. Matern Child Hlth J, 2013; 17:1151–1157.

15.

Mackay

, Ritthiphakdee

, Reddy

. Tobacco control in Asia. Lancet, 2013; 381:1581–1587.

16.

Tarrant

, Lok

KYW

, Fong

DYT

, et al. Effect of a hospital policy of not accepting free infant formula on in-hospital formula supplementation rates and breastfeeding duration. Public Health Nutr, 2015; 18:2689–2699.

17.

Baby Friendly Hospital Initiative Hong Kong Association. World breastfeeding week 2017: annual summary, 2017.

18.

World Health Organization. Indicators for Assessing Breast-Feeding Practices. Geneva: WHO. 1991; 91:14.

19.

Labbok

, Krasovec

. Toward consistency in breastfeeding definitions. Stud Fam Plann, 1990; 21:226–230.

20.

Tarrant

, Fong

, Wu

, et al. Breastfeeding and weaning practices among Hong Kong mothers: A prospective study. BMC Pregnancy Childbirth, 2010; 10:27.

21.

Thulier

, Mercer

. Variables associated with breastfeeding duration. J Obstet Gynecol Neonatal Nurs, 2009; 38:259–268.

22.

Stata Corp. Stata Statistical Software: Release 14.1 [program] [computer program]. College Station, StataCorp LP, 2015.

23.

, Binns

, Zhang

, et al. Paternal smoking and breastfeeding in Xinjiang, PR China. J Hum Lact, 2010; 26:242–247.

24.

Andersen

, Ronn

, Tjonneland

, et al. Low maternal but normal fetal prolactin levels in cigarette smoking pregnant women. Acta Obstet Gynecol Scand, 1984; 63:237–239.

25.

Baron

, Bulbrook

, Wang

, et al. Cigarette smoking and prolactin in women. BMJ, 1986; 293:482–483.

26.

Amir

. Maternal smoking and reduced duration of breastfeeding: A review of possible mechanisms. Early Hum Dev, 2001; 64:45–67.

27.

Bai

, Fong

, Lok

KYW

, et al. Relationship between the Infant Feeding Preferences of Chinese mother's immediate social network and early breastfeeding cessation. J Hum Lact, 2016; 32:301–308.

28.

Odom

, Li

, Scanlon

, et al. Association of family and health care provider opinion on infant feeding with mother's breastfeeding decision. J Acad Nutr Diet, 2014; 114:1203–1207.

29.

Natland

, Andersen

, Nilsen

TIL

, et al. Maternal recall of breastfeeding duration twenty years after delivery. BMC Med Res Methodol, 2012; 12:179.

30.

Curtin

, Morabia

, Bernstein

. Relation of environmental tobacco smoke to diet and health habits: Variations according to the site of exposure. J Clin Epidemiol, 1999; 52:1055–1062.

31.

Horta

, Victora

, Menezes

, et al. Environmental tobacco smoke and breastfeeding duration. Am J Epidemiol, 1997; 146:128–133.

32.

Leung

, Ho

, Lam

. Secondhand smoke exposure, smoking hygiene, and hospitalization in the first 18 months of life. Arch Pediatr Adolesc Med, 2004; 158:687–693.

33.

Matt

, Quintana

, Destaillats

, et al. Thirdhand tobacco smoke: Emerging evidence and arguments for a multidisciplinary research agenda. Environ Health Perspect, 2011; 119:1218–1226.

34.

Leung

, Ho

, Wang

, et al. Secondhand smoke from multiple sources, thirdhand smoke and respiratory symptoms in Hong Kong adolescents. Nicotine Tob Res, 2018; 20:192–198.