The association between secondhand smoke exposure and growth outcomes of children: A systematic literature review

INTRODUCTION The strong relation between maternal smoking and maternal secondhand smoke (SHS) exposure and the growth of newborn infants has been proven. However, the effect of SHS on growth outcomes of older children is not well defined. Through a systematic literature review, we sought to determine whether a relationship exists between SHS exposure and growth outcomes of children up to 8 years of age. METHODS A systematic review was performed, including articles published between 2004–2019, related to SHS exposure (prenatal and postnatal) and children’s growth (weight, length/height, and head circumference). The relevant articles were identified from Science Direct, ProQuest, Sage Publication, Scopus, Wiley Online Library, CINAHL Plus with Full Text (via EBSCOhost) and Google search. RESULTS Seventeen articles were identified, of which three categories of growth measurements were extracted, comprising weight (weight, WAZ, WHZ, and BMI), height (height/length and HAZ) and head circumference. SHS exposure both pre or postnatally was inversely associated with weight (deficit in weight, risk of underweight, risk of wasting) and height (lower length and risk of stunting) and elevated BMI of children. Furthermore, prenatal SHS exposure was associated with a lower head circumference. CONCLUSIONS The current review identified that exposure to SHS may be associated with adverse growth outcomes in children. It is crucial that active smokers, specifically those who live with children or with a pregnant partner, are made aware of the potential effects of SHS exposure on non-smokers. Further assessment of the association between exposure to SHS and other growth outcomes in other age groups is needed.


INTRODUCTION
More than a third of the global population are passive smokers and regularly exposed to the dangerous effects of tobacco smoke. Smoke exposure is responsible for approximately 0.6 million deaths annually and approximately 1% of global disease around the world 1 . The result of a study across 192 countries showed that 40% of children were exposed to secondhand smoke (SHS) 2 and 36% were exposed to SHS in utero 3 . This makes the implications of exposure a potentially significant public health problem.
Early childhood (usually defined as a newborn baby until the age of 8 years) is the phase of incredible growth in several aspects: physical, cognitive, socialemotional, and language skills 4,5 . During the early years, the brain develops quickly and has a high capacity for change, with the foundation set for health and wellbeing throughout life. Therefore, this period is critical. Protecting children from threat, including secondhand smoke exposure, is part of nurturing care that is sensitive to children's health and nutrition needs 5 . The existing studies showed that SHS exposure has a strong relation with low birth weight [6][7][8] , premature birth 9 , shorter baby length 10 , higher risk of fetal death, congenital defects 11 , and childhood obesity 12 . To date, limited information has been collated to illustrate the association between SHS exposure in non-smoking mothers during pregnancy and/or in children during postnatal life and the growth of children. Two review studies examined the association between SHS exposure and children growth outcomes, focusing on tobacco use of the mother during pregnancy 13,14 . Two other review articles explored the impact of SHS exposure on non-smoking pregnant women on anthropometric growth of children, focusing on the newborn baby 15,16 .

Review Paper
The objective of this systematic literature review was therefore to determine whether SHS exposure was associated with growth outcomes in children up to 8 years of age.

Search strategy
We identified the eligible literature through a systematic search in 7 electronic databases: Science Direct, ProQuest, Sage Publication, Scopus, Wiley Online Library, CINAHL Plus with Full Text (via EBSCOhost) and Google search, without timewindow restriction. The search used a combination of keywords from SHS (tobacco, tobacco smoke, environmental tobacco smoke, passive smoking, and secondhand smoke) and growth (anthropometric, growth, weight, length, and head circumference). After the abstracts were retrieved and screened, we evaluated the full text of the articles that related to SHS exposure and children's growth. The additional articles were searched using the bibliography of the selected articles. The literature search was completed in June 2019.

Inclusion and exclusion criteria
To be eligible for inclusion, the article must present the data from an observational study that includes both a measure of SHS exposure (pre or postnatal) and that at least one of the research objectives is measuring the children's growth through anthropometric measurements. The anthropometry indices were weight, height, head circumference, weight-for-age z-score (WAZ), length or height-for-age z-score (HAZ), weight-for-length or weight-forheight z-score (WHZ), head circumference z-score (HCZ), body mass index-for-age z-score (BMIZ) or BMI-for-age percentile, and BMI or Kaup index. BMI and Kaup indices divide weight by the square of the height (kg/m 2 ) 17,18 . Underweight (< -2 SD WAZ), stunting (< -2 SD HAZ), wasting (< -2 SD WHZ and BMIZ) and overweight (> +2 SD WHZ and BMIZ; ≥85 percentile BMI-for-age) are used to measure nutritional imbalance resulting in malnutrition (assessed from underweight, wasting, and stunting) and overweight [19][20][21] . Also, Z scores < -3 SD for WHZ, WAZ or HAZ were considered as severely wasted, severely underweight, or severely stunted 22 .
Retrieved articles were excluded if the exposure and outcome variables were not defined clearly or if the association of the growth outcome with SHS exposure could not be determined independently of other toxins such as air pollution or illicit drug exposure in utero, due to these factors being combined into one variable. This paper focused on SHS exposure on children aged <8 years. Therefore, if an article only included children aged ≥8 years, the article was excluded.
In addition, if no statistical evidence relevant to our research question was presented (e.g. data not shown) or were not original research articles, these were also excluded 23 . As the effect of maternal smoking in the prenatal period on the growth of the offspring has been reviewed thoroughly elsewhere 13,14 , the aim of the present review was to emphasize SHS exposure from other people smoking (i.e. paternal smoking). Therefore, articles that had data only on maternal smoking in pregnancy were excluded from the present systematic literature review.

Data quality assessment
The quality of the studies was appraised using a scale adapted from the Newcastle/Ottawa Scale (NOS) (the appraisal standard of NOS is presented in the Supplementary file). Each study was assessed using the point system based on the NOS. One point was added when a study included relevant information that could be related to the NOS. Eight items in cohort studies and five items in cross-sectional studies that could be related to the NOS were identified. Hence, cross-sectional studies assigned 5, 4, 3, or 0-2 points, were assessed as 'very good', 'good', 'satisfactory' or 'unsatisfactory', respectively. Also, cohort studies with 7-8, 5-6, 4, or 0-3 points, were classified as 'very good', 'good', 'satisfactory' or 'unsatisfactory', respectively. Unsatisfactory studies were excluded 24,25 .

Data extraction
Any issues that occurred are discussed below for each article. For all articles, the following data were independently extracted: year of publication, study design, participant sampling, country, number of participants, mean participant age, participant gender, percentage of participants exposed to SHS, SHS measurement, growth measurement, covariates included in the analysis, and the study outcome.

RESULTS
The literature search identified 2138 records, of which 2105 were excluded after screening title and abstract for relevance to the research question and removal of duplicates. Using the inclusion and exclusion criteria, we selected 34 full texts for further assessment. A total of 17 studies were excluded after screening for relevant and sufficient information, including maternal smoking during pregnancy only, no statistical evidence, not using an anthropometric measure, not assessing relationship under question, and anthropometric measurement in adults only. In total, 17 studies were included in our final systematic review ( Figure 1).
The articles were published between 2004-2019, eight were cross-sectional 21,[26][27][28][29][30][31][32] , and the other nine studies were prospective studies 17,18,[33][34][35][36][37][38][39] . The age range included in this review was from 0 to 7 years old. Nine studies were conducted in Asian countries, two in the USA, four in Europe, and the last two 29,32 were conducted in multi-countries (18 and 7 countries, respectively). Four articles used a biomarker for SHS exposure assessment, including cotinine serum 17 , plasma cotinine 18 , and urinary cotinine 34,37 . Meanwhile, the other articles used an interview or self-report through a questionnaire, from parents or caregivers, to estimate the SHS exposure. The articles are presented in two sections: growth measurements in children exposed to prenatal SHS (n=5; Table 1), and growth measurements in children exposed to postnatal SHS (n=12; Table 2).
From 17 articles selected, the current review explored three anthropometric measurements as the outcome of SHS exposure from each of 8 studies, two anthropometric measurements were captured from each of 2 studies, and only one anthropometric measurement was taken from each of 7 articles. The present review then classified measurements into three groups, comprising weight (weight, WAZ, WHZ and BMI), height (height/length and HAZ), and head circumference.

SHS exposure and weight of children
Nine studies investigated the effect of SHS exposure on at least one measurement of weight, WAZ or WHZ. While seven studies explored the association between SHS exposure and at least one measurement of BMI (BMI, BMIZ, BMI-for-age percentile or Kaup index). Exposure to SHS was inversely associated with weight outcome (deficit in weight, risk of underweight, risk of wasting) in 7 of 9 studies [26][27][28]30,31,33,37 . The remaining two studies presented no association between them 35,39 . Only two of nine studies evaluated exposure of SHS during the prenatal period. Furthermore, seven studies conducted in low-income and lower middle-income countries and two other studies performed in upper middle-income countries based on World Bank classification 2019-2020 40 .
SHS exposure was associated with higher BMI or overweight in 4 of 7 studies 17,21,34,38 . One study showed no association 36 and one study presented an inverse association between SHS exposure and BMI 18 . The last study by Braithwaite et al. 29 revealed two contrasting results in high and low GNI (gross national income      per capita) countries, with SHS exposure associated with higher BMI in high GNI countries and lower BMI in low GNI countries. Three of seven studies, conducted in children aged 6-7 years, found higher BMI in exposed children 21,29,38 . Also, all studies in high-income and upper middle-income countries, except one, were conducted in 18 countries with two levels of GNI.

SHS exposure and height of children
Ten studies examined the effect of SHS exposure on at least one height indicator (height/length and HAZ), from those two studies conducted in the prenatal period. Eight of ten studies were performed in low-income and lower middle-income countries and the other two were in upper middle-income countries. SHS exposure was associated with lower length in two studies 33,37 and a higher risk of stunting in four studies 26,28,30,31 . Nevertheless, exposure to SHS and length/height or risk of stunting were not related in four other studies 27,32,35,39 .

SHS exposure and head circumference
Three studies evaluated the effect of SHS exposure on the head circumference (HC) in children. Two studies, performed during the prenatal period, found a significant association between SHS exposure and lower HC in children 33,39 . Finally, one study was conducted Cohort of Belarusian children 13889 Children 6.5 years of age; 47.% female; 51.2% exposed to SHS Self-reported by mother (exposed vs not exposed)  for the postnatal period, and the head circumference was found not to be significantly different between the exposed and non-exposed to SHS 37 .

DISCUSSION
This review notes that SHS exposure during the pre or postnatal period has adverse effects on weight and height outcomes in childhood. There is also evidence that SHS exposure in the prenatal period is associated with a lower head circumference. There are several potential mechanisms on how prenatal exposure influences growth in children. SHS contains more than 4000 chemical substances among which are some of the main carcinogenic substances, such as Polycyclic aromatic hydrocarbons (PAHs), 4-aminobiphenyl (ABP), tobacco-specific nitrosamines N'nitrosonornicotine (NNN), and 4-(methylnitrosamino-) 1-(3,pyridyl)-1-butone (NNK) 41 . PAHs, ABP and N-nitrosamines may cross from the maternal serum to fetus circulation [42][43][44] . In passive smoker mothers, PAHs and NNK might pass through the placenta and directly influence the children's hypothalamic centres, which may delay body growth 18 . It is known that the hypothalamus has a vital function in the control of body weight by balancing food intake, energy release, and body fat storage 45 . Moreover, a study showed that height growth of children exposed to cigarette smoke was lower because the smoke contains cadmium, which disturbs zinc bioavailability 46 . PAHs and NNK may also go through the placenta and directly influence the volume of the fetus anterior cingulate region, and this condition may cause a lower head circumference of the baby 18 . Head growth during prenatal period and infancy is crucial as it is related to subsequent IQ development and is essential in determining how well cognitive abilities are maintained in old age 47,48 .
Another reason might be related to lower nutrition in SHS exposed children, due to family income spent on cigarettes rather than food 26,28,30,49 . Furthermore, SHS causes frequent health problems in infants and children 50 . Based on UNICEF's conceptual framework on child undernutrition, inadequate dietary intake and frequent illness are immediate causes of child undernutrition 51 . A study by Danaei et al. 52 , in 137 developing countries, demonstrated that fetal growth restriction (FGR) and bad sanitation were the leading risk factors for stunting in developing countries.
Passive smoking during pregnancy is notably associated with an increased incidence of FGR. The present review also reveals the association between parental smoking and child stunted growth.
The present review also showed an association between prenatal or postnatal SHS exposure and higher BMI, particularly in children aged 6-7 years. A study by Braun et al. 17 found stronger effects of tobacco smoke exposure as children become older. Our review is in line with a meta-analysis by Oken et al. 13 on 14 articles (with 84563 children) and Magalhães et al. 53 that children whose mothers smoked during the prenatal period were at an elevated risk of becoming overweight in childhood (OR=1.5 and OR=1.43, respectively). A meta-analysis by Qureshi et al. 54 demonstrated the association between prenatal exposure to environmental tobacco smoke and childhood obesity with OR=1.905.
Prenatal SHS exposure of the mother might cause low birth weight (LBW). It might lead to LBW through the potential pathways of maternal inflammation and lower placental weight 6 . LBW is a proxy-marker of poor fetal growth and nutrition. Based on the Developmental Origins of Health and Disease (DOHaD) hypothesis, the underlying mechanism is poor nutrition (it might be due to nicotine exposure) in utero or during early childhood that affects the risk of disease later in life. Some of the mechanisms begin at the time of the perinatal insult, while other mechanisms perform a more significant part in influencing metabolic disease during the postnatal period (i.e. during catch-up growth). It is similar to the concepts of fetal programming and Barker's hypothesis, which illustrate the relationship between a specific path of growth-consisting of slow growth in utero and rapidly increasing BMI in postnatal period-and the development of chronic diseases later in life, such as coronary heart disease and related disorders including stroke, hypertension and non-insulin dependent diabetes 55,56 .
Both undernutrition and overnutrition have similar long-lasting physiologic effects. Undernutrition increases susceptibility to fat accumulation, insulin resistance in adulthood, hypertension, dyslipidaemia and a reduced capacity for manual work, among other impairments 57 . Elevated BMI in childhood predicts risk of hypertension in young adulthood, type 2 diabetes, and, to a lesser extent, cardiovascular diseases 58,59 .

Strengths and limitations
The strengths of this review include its wide-ranging search strategy, systematic data extraction and quality assessment method used. However, there are some limitations. These include the number of participants among extracted articles, the relatively significant difference between study areas, and limited to South-Eastern and Western Asia. These factors might affect the results of the review. At the same time, it reveals the need to further investigate the association between secondhand smoke exposure and growth measurement of children in other countries. Furthermore, the small number of published studies, particularly on head circumference, as an outcome of SHS exposure, prevents us from drawing firm conclusions.

CONCLUSIONS
The current review emphasizes that growth (below or above the standard) in children may be affected by secondhand smoke exposure pre or postnatally. SHS exposure should thus be considered a modifiable risk factor for underweight, wasting and stunting, specifically in low-income and lower middle-income countries; elevated BMI and overweight particularly in high-income and upper middle-income countries; and small head circumference that might be due to prenatal SHS exposure.
This review implies that it is crucial that people who currently are active smokers, specifically those who live with children or with a pregnant partner, are made aware of the potential effect of tobacco smoke exposure on non-smokers. By encouraging household members to stop smoking (and/or by declining smoking prevalence rates in the population as a whole), the burden of children's growth problems would also be reduced at the population level. Furthermore, it is also important to encourage families to maintain a smoke-free home environment, and hence education on the health risks of SHS exposure may protect non-smoking women and their children from SHS exposure and its potential negative effects on growth outcomes.