Maternal smoking behaviour across the first two pregnancies and small for gestational age birth: Analysis of the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the South of England

Elizabeth J Taylor; Pia Doh; Nida Ziauddeen; Keith M Godfrey; Ann Berrington; Nisreen A Alwan

doi:10.1371/journal.pone.0260134

. 2021 Nov 18;16(11):e0260134. doi: 10.1371/journal.pone.0260134

Maternal smoking behaviour across the first two pregnancies and small for gestational age birth: Analysis of the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the South of England

Elizabeth J Taylor ^1,^2,^*, Pia Doh ¹, Nida Ziauddeen ¹, Keith M Godfrey ^2,³, Ann Berrington ⁴, Nisreen A Alwan ^1,^2,^5,^*

Editor: JJ Cray Jr⁶

¹School of Primary Care, Population Sciences and Medical Education, Faculty of Medicine, University of Southampton, Southampton, United Kingdom

²NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, United Kingdom

³MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton General Hospital, Southampton, United Kingdom

⁴Department of Social Statistics and Demography, University of Southampton, Southampton, United Kingdom

⁵NIHR Applied Research Collaboration (ARC) Wessex, Southampton, United Kingdom

⁶Ohio State University, UNITED STATES

Competing Interests: The authors have read the journal’s policy and the authors of this manuscript have the following competing interests: EJT is supported by an NIHR Southampton Biomedical Research Centre and University of Southampton Primary Care, Population Sciences and Medical Education PhD studentship. Within 5 years of conducting this study, NAA has received research grants as PI from the Academy of Medical Sciences/Wellcome Trust and the National Institute for Health Research (NIHR) Southampton Biomedical Research Centre and the NIHR Applied Research Collaboration, Wessex. NAA has acted as a member of the National Institute for Health and Care Excellence (NICE) Antenatal Care Guideline Committee. KMG has received reimbursement for speaking at conferences sponsored by companies selling nutritional products and is part of an academic consortium that has received research funding from BenevolentAI Bio Ltd, Abbott Nutrition, Nestec and Danone. KMG is supported by the UK Medical Research Council (MC_UU_12011/4), the National Institute for Health Research (NIHR Senior Investigator (NF-SI-0515-10042), NIHR Southampton 1000DaysPlus Global Nutrition Research Group (17/63/154) and NIHR Southampton Biomedical Research Centre (IS-BRC-1215-20004)), the European Union (Erasmus+ Programme Early Nutrition eAcademy Southeast Asia-573651-EPP-1-2016-1-DE-EPPKA2-CBHE-JP and ImpENSA 598488-EPP-1-2018-1-DE-EPPKA2-CBHE-JP), the British Heart Foundation (RG/15/17/3174), the US National Institute On Aging of the National Institutes of Health (Award No. U24AG047867) and the UK ESRC and BBSRC (Award No. ES/M00919X/1). AB has received funding from the European Commission (Horizon 2020), UKRI(ESRC) and the National Institute for Health Research. The other authors have no potentially competing interests to declare. This does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials.

^✉

* E-mail: E.J.Taylor@soton.ac.uk (EJT); N.A.Alwan@soton.ac.uk (NAA)

Roles

Elizabeth J Taylor: Formal analysis, Methodology, Visualization, Writing – original draft, Writing – review & editing

Pia Doh: Writing – original draft, Writing – review & editing

Nida Ziauddeen: Data curation, Writing – review & editing

Keith M Godfrey: Supervision, Writing – review & editing

Ann Berrington: Supervision, Writing – review & editing

Nisreen A Alwan: Conceptualization, Formal analysis, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing

JJ Cray Jr: Editor

PMCID: PMC8601508 PMID: 34793557

Abstract

Maternal smoking is established to cause adverse birth outcomes, but evidence considering maternal smoking change across successive pregnancies is sparse. We examined the association between self-reported maternal smoking during and between the first two pregnancies with the odds of small for gestational age (SGA) birth (<10^th percentile) in the second infant.

Records for the first two pregnancies for 16791 women within the SLOPE (Studying Lifecourse Obesity PrEdictors) study were analysed. This is a population-based cohort of prospectively collected anonymised antenatal and birth healthcare data (2003–2018) in Hampshire, UK. Logistic regression was used to relate maternal smoking change to the odds of SGA birth in the second infant.

In the full sample, compared to never smokers, mothers smoking at the start of the first pregnancy had higher odds of SGA birth in the second pregnancy even where they stopped smoking before the first antenatal appointment for the second pregnancy (adjusted odds ratio (aOR) 1.50 [95% confidence interval 1.10, 2.03]). If a mother was not a smoker at the first antenatal appointment for either her first or her second pregnancy, but smoked later in her first pregnancy or between pregnancies, there was no evidence of increased risk of SGA birth in the second pregnancy compared to never smokers. A mother who smoked ten or more cigarettes a day at the start of both of her first two pregnancies had the highest odds of SGA birth (3.54 [2.55, 4.92]). Women who were not smoking at the start of the first pregnancy but who subsequently resumed/began smoking and smoked at the start of their second pregnancy, also had higher odds (2.11 [1.51, 2.95]) than never smokers.

Smoking in the first pregnancy was associated with SGA birth in the second pregnancy, even if the mother quit by the confirmation of her second pregnancy.

Introduction

Maternal smoking has been associated with the inability to conceive as well as the risks of ectopic pregnancy, miscarriage, stillbirth and prematurity [1, 2] and the association between smoking during pregnancy and fetal growth restriction is considered to be causal [1]. A dose response relationship has been shown between the number of cigarettes smoked a day in pregnancy and the risk of placental abruption and negative birth outcomes [1, 3, 4]. The greatest morphological effects in the placenta are found where there is heavy smoking before 10 weeks gestation (> 20 cigarettes a day) [2]. In addition to being born prematurely [5], adverse health consequences for the child include being born small for gestational age (SGA) (<10^th percentile) [6] and an increased risk of congenital malformations, primarily oral-facial clefts [7].

A recent systematic review and meta-analysis has estimated that nearly 2% of women globally smoke during pregnancy, with nearly three-quarters of these smoking daily [8]. There is substantial variation between the countries considered in this study with the highest estimated prevalence being in Ireland (38.4% [95% CI [25.4, 52.4]), Uruguay (29.7% [16.6, 44.8]) and Bulgaria (29.4% [26.6, 32.2]) [8]. Figures for the third quarter of 2019/20 show that in England, where this study is based, 10.5% of women report smoking at the time of delivery, although there is substantial regional variation between the lowest and highest rates (from 1.6% in Central London to 23.3% in Blackpool) [9].

Since longitudinal data are sparse, most studies are only able to consider the association between maternal exposures, such as smoking, in one pregnancy with the outcome for that pregnancy, and biological links during the same pregnancy are already established. Few studies have sought to categorise maternal smoking behaviour across successive pregnancies to examine whether the association between SGA and history of smoking extends beyond the period of the same pregnancy or whether exposure in a previous pregnancy, or during the interconception period also carries risk of having a SGA birth in a subsequent pregnancy.

Changes to DNA methylation patterns have been seen in the placentas of women who quit smoking prior to pregnancy and a recent study suggests that tobacco exposure may cause long-term effects via the transmission of epigenetic marks to non-directly exposed placentas [10]. A narrative review of epigenetic alterations due to maternal tobacco smoking in pregnancy concluded that there is increasing evidence to indicate that such alterations persist postnatally, but that there is also the suggestion of some reversibility of DNA methylation when stopping smoking either before or during pregnancy [11].

An analysis of Norwegian Medical Birth Registry data (1999 to 2014) found that daily smoking throughout both of the first two pregnancies was associated with nearly three times the risk of the second child being born SGA (compared to non-smokers in both pregnancies), but that quitting before or during the second pregnancy reduced the risk [12].

We aimed to characterise maternal smoking behaviours across a mother’s first two pregnancies and examine the relation of smoking behaviours with the second child’s risk of being born SGA. In doing so we examine the hypothesis that mothers who smoked in a previous pregnancy or who smoked between pregnancies have a higher risk of SGA in the second pregnancy compared to never smokers, even if they were not smoking during the second pregnancy. Associations could potentially arise through a variety of biological mechanisms, and these include the effects of smoking on nutritional status or periconceptional development [13, 14]. Whether such a link is biological or not would depend on how much is it confounded by other factors. This study is observational and so we cannot establish causality, however we believe if such associations were demonstrated this would open the way to exploring possible causal mechanisms.

The exposure groups to be examined include mothers who smoked in their first pregnancies but who quit smoking at some point up to the confirmation of the second pregnancy and those who initiated or resumed smoking after the first antenatal appointment (ANA) for their first pregnancy and reported smoking at the first ANA for their second pregnancy. We also examined non-smokers at the start of both pregnancies but with a history of smoking before one or both pregnancies. Hence, our comparison group was those who never smoked. Identifying women in these groups may enable the targeting of women for interventions.

In addition, we wanted to explore if these relationships are different based on previous history of SGA in the first pregnancy.

Methods

The SLOPE (Studying Lifecourse Obesity PrEdictors) study is a population-based anonymised cohort of prospectively collected routine antenatal healthcare data collected between January 2003 and April 2018 for women registered with University Hospital Southampton NHS Trust Maternity Services, Hampshire, UK [15–17]. Records for 16791 women with their first two consecutive singleton live-birth pregnancies were included (Fig 1).

Exclusions from the data are detailed in Fig 1. Births which took place before 24 weeks or after 42 weeks gestation were excluded as SGA reference values do not exist for these gestations. An exclusion for pregnancies where the first ANA for the second pregnancy took place after 168 days gestation (as assessed by ultrasound examination performed by healthcare professionals) was made since these were likely to be high-risk pregnancies referred from elsewhere. Variables documenting the previous numbers of live and stillbirths were used to identify women giving birth for the first and second time and to exclude women who either had a first or second birth elsewhere or who had a stillbirth prior to their first live birth or between live births.

This analysis forms part of a research project approved by the University of Southampton Faculty of Medicine Ethics Committee (ID 24433) and the National Health Service Health Research Authority (IRAS 242031).

Assessment of the exposure

Self-reported smoking status was recorded by a midwife at the first ANA for each pregnancy. For an uncomplicated pregnancy this is recommended to take place by 10 weeks gestation [18]. Women were asked to self-report smoking status at this appointment, and were asked if they were current smokers or if they had ever smoked. If they reported being a current smoker, they were asked how many cigarettes a day they smoked (up to 10 a day/between 10 and 20 a day/more than 20 a day) and the response recorded. Those who reported that they were ex-smokers were asked when they stopped smoking (more than 12 months before conception/less than 12 months before conception/on confirmation of the current pregnancy).

Exposure category definitions

A variable was derived to characterise smoking behaviour across the first two pregnancies based on the responses given at the first ANAs for each pregnancy. The full derivation of this variable is given in Table 1.

Table 1. Summary of derived smoking categories based on self-reported maternal smoking status recorded at the first antenatal appointment for each pregnancy.

Derived smoking category	Smoking status recorded at first ANA for P1	Smoking status recorded at first ANA for P2	Additional notes
Heavier smoker	Smoking 10 or more cigarettes a day	Smoking 10 or more cigarettes a day	These women are the heaviest smokers at the start of each pregnancy
Smoker	Smoking up to 10 cigarettes a day	Smoking up to 10 cigarettes a day
Smoker increased	Smoking up to 10 cigarettes a day	Smoking 10 or more cigarettes a day	These women report an increase in the number of cigarettes smoked from the first ANA of P1 to the first ANA of P2
Smoker reduced	Smoking 10 or more cigarettes a day	Smoking up to 10 cigarettes a day	These women report a reduction in the number of cigarettes smoked from the first ANA of P1 to the first ANA of P2
Smoker P2 (not smoking at the first ANA P1)	Not smoking. May be an ex-smoker or have never smoked. If an ex-smoker may have quit at any point up to the confirmation of P1	Smoking any number of cigarettes	These women may have initiated or resumed smoking at any point after the first ANA for P1
Smoker P1 (stopped before the first ANA P2)	Smoking any number of cigarettes	An ex-smoker who quit at any point up to the confirmation of P2	These women may have quit smoking at any point after the first ANA for P1; the latest point for cessation would have been on the confirmation of P2
Other smoker (smoker later in P1 or between pregnancies; not smoking at first ANA for P1 or P2)	A non-smoker or an ex-smoker who quit at any point before P1 conception or on confirmation of P1	An ex-smoker who quit either less than 12 months before P2 conception or on confirmation of P2	These women did not report smoking at the first ANA for either P1 or P2. They could have smoked later in P1 or after the birth of their first child. They will have smoked at some stage during the 12 months prior to the conception of their second child
Ex-smoker	An ex-smoker who quit at any point up to the confirmation of P1	An ex-smoker who quit more than 12 months before the conception of P2	These women may have smoked after the first ANA for P1 but did not smoke during the 12 months prior to the conception of their second child
Never smoker	Non-smoker with no past history of smoking	Non-smoker with no past history of smoking

Open in a new tab

Abbreviations: ANA, antenatal appointment; P1, first pregnancy; P2 second pregnancy.

Outcome assessment

Age and sex-specific birth weight centiles were used to classify infants born SGA [19]. This was defined as < 10^th percentile. Baby’s birthweight (grams) was measured and sex was recorded at birth as part of routine care by a healthcare professional. Gestational age (days) was calculated based on a first trimester ultrasound dating scan [18].

Assessment of covariates

Maternal age (in years) was calculated from date of birth prior to the extraction of the dataset. Maternal weight was measured by a midwife at the first ANA for each pregnancy (kilograms). Height was self-reported (metres) and body mass index (BMI) was then derived (kg/m²). Self-reported variables collected at the first booking appointment for each pregnancy included maternal ethnicity, highest level of educational attainment (secondary (GCSEs) or below/college (A levels)/university degree or above), employment status (condensed to yes/no), partnership status (partnered/lone parent), folic acid supplementation (taking prior to pregnancy/at confirmation of pregnancy/not taking) and infertility treatment (condensed to yes/no). Gestational diabetes mellitus (GDM) and gestational hypertension were identified later during each pregnancy and the diagnosis reported in the database. The interpregnancy interval (days) was calculated based on the World Health Organisation definition [20] by taking the period from the date of the first birth to the conception of the second birth, using the gestational age of the second child. SGA in the first pregnancy was calculated as described in the outcome assessments section above.

Statistical analysis

Unadjusted comparisons were carried out using Chi-squared tests for categorical variables and ANOVA for continuous variables.

The association between change in smoking behaviour between pregnancies and the risk of SGA birth in the second pregnancy was examined by fitting logistic regression models predicting a binary outcome (SGA/not SGA). A minimal sufficient adjustment set of confounding variables was identified using a directed acyclic graph (DAG) constructed using DAGitty.net [21, 22] (Fig 2). The DAG illustrates the hypothesised confounding relationships by factors collected at the start of each pregnancy and explicitly identifies our assumptions using a priori causal knowledge [21, 23]. References to maternal education and employment in the DAG are taken to be those recorded at the start of the first pregnancy in our analysis.

A large number of minimal sufficient adjustment sets were identified using DAGitty.net [21, 22]. We selected a parsimonious set comprising maternal age, BMI, educational attainment, employment status, partnership status, folate supplementation and infertility treatment details collected at the start of the first pregnancy, diagnoses of gestational diabetes mellitus and gestational hypertension recorded during the first pregnancy, SGA birth in the first pregnancy, maternal ethnicity and the length of the interpregnancy interval (Model 1). The variables were complete in all but 72 cases. In 551 cases ethnicity was not recorded and has been included as “Not specified”.

Each minimal adjustment set identified should close all biasing paths, leaving only measured causal paths open [24]. We used the other sets identified, some of which included covariates collected at the start of or during the second pregnancy, to confirm that there was no change to the results of our analyses and this sensitivity analysis is presented in S1 Table.

Whilst the minimal adjustment set used in this analysis consists mainly of covariates identified at the start of the first pregnancy, a number of second pregnancy covariates may be mediators of the effect of interpregnancy smoking change on SGA birth in the second pregnancy. Analysis was also therefore undertaken to take account of potential mediators. This analysis also followed a minimal adjustment set identified by DAGitty.net [22], but this time taking account of mediators. The identified adjustment set was the same as that identified for Model 1, but with gestational diabetes mellitus and gestational hypertension diagnosed in the second pregnancy in place of that for the first pregnancy and with the addition of maternal BMI recorded at the start of the second pregnancy (Model 2). The adjustment set for Model 2 should close all other measured causal paths with the exception of the effect of interpregnancy smoking change on SGA birth in the second pregnancy [24].

For each Model, analysis was initially undertaken in the whole sample and was then stratified to examine the association with new SGA (where there was no SGA birth in the first pregnancy) and recurrent SGA (following SGA birth in the first pregnancy). Stratified analysis was undertaken on this basis since women who have had a previous SGA birth are known to be at higher risk for a subsequent SGA birth, and therefore previous SGA is hypothesised to be an effect modifier of the effect of smoking on the probability of second SGA [25]. We aimed to assess if the effect estimates are different for the risk of recurrent SGA and new SGA. Our comparison group for all our analyses was never smokers.

All analysis was performed using R [26]. Packages used included data.table [27], dplyr [28], epiDisplay [29], ggplot2 [30], haven [31], psych [32], reshape [33] and tidyr [34].

Results

Maternal and infant socio-demographics in the second pregnancy, categorised by exposure, are given in Table 2. Of the 16791 women included in this analysis, 49.9% (n = 8386) were categorised as never smokers. There was a slight reduction in the overall percentage of women who reported smoking at the first antenatal appointment for the first pregnancy (15.0%) and the first antenatal appointment for the second pregnancy (13.3%).

Table 2. Maternal characteristics recorded at the first antenatal appointment at the start of the second pregnancy, together with characteristics of both the first and second infants.

	Never smoker	Heavier smoker	Smoker	Smoker increased	Smoker reduced	Smoker P2¹	Smoker P1²	Other smoker³	Ex-smoker	p-value⁴
n	8386	333	791	347	313	456	738	1347	4080
Age, years	30.2 (5.0)	23.8 (4.4)	24.6 (4.8)	23.4 (3.9)	24.2 (4.6)	24.7 (4.8)	25.7 (4.7)	27.1 (5.2)	30.1 (5.1)	< 0.001
(mean, SD)	30.2 (5.0)	23.8 (4.4)	24.6 (4.8)	23.4 (3.9)	24.2 (4.6)	24.7 (4.8)	25.7 (4.7)	27.1 (5.2)	30.1 (5.1)	< 0.001
Timing of ANA, weeks (mean, SD)	11.0 (2.3)	11.3 (3.3)	11.2 (2.9)	11.4 (3.3)	10.9 (2.6)	11.0 (2.9)	11.0 (2.6)	10.8 (2.5)	11.0 (2.2)	0.001
BMI, kg/m²	25.3 (5.3)	26.1 (6.4)	25.8 (6.1)	26.7 (6.5)	26.5 (6.5)	26.6 (6.3)	26.8 (6.1)	26.7 (6.0)	26.2 (5.6)	< 0.001
(mean, SD)	25.3 (5.3)	26.1 (6.4)	25.8 (6.1)	26.7 (6.5)	26.5 (6.5)	26.6 (6.3)	26.8 (6.1)	26.7 (6.0)	26.2 (5.6)	< 0.001
Length of IPI, weeks (median, IQR)	96	91	107	98	113	121	130	123	96	< 0.001
Length of IPI, weeks (median, IQR)	(63, 144)	(46, 164)	(58, 184)	(52, 163)	(58, 189)	(68, 188)	(74, 217)	(74, 190)	(62, 147)	< 0.001
BMI category:										< 0.001
Underweight	2.9	6.6	4.9	4.9	6.1	5.0	2.2	1.8	1.8
Normal weight	54.9	42.9	48.4	42.1	42.5	44.5	45.0	44.0	48.6
Overweight	25.8	26.1	24.4	25.4	26.2	23.0	27.8	29.3	29.3
Obese	16.4	24.3	22.3	27.7	25.2	27.4	25.1	24.9	20.3
Ethnicity:										< 0.001
White	79.4	97.9	94.1	96.8	96.5	94.5	95.3	93.5	92.6
Other ethnicities	17.1	0.6	2.8	1.7	1.3	2.6	3.0	3.6	3.6
Not specified	3.5	1.5	3.2	1.4	2.2	2.9	1.8	2.8	3.7
Highest education level:										< 0.001
University or above	45.8	1.5	4.8	3.2	5.4	5.9	7.6	15.1	34.2
College	34.9	39.6	48.3	46.1	46.3	53.3	53.4	52.3	44.9
Secondary or below	19.2	58.9	46.9	50.7	48.2	40.8	39.0	32.7	20.9
In employment	72.3	28.9	41.7	33.3	38.6	44.6	55.4	64.7	73.8	< 0.001
(missing records)	(n = 57)	(n = 1)	(n = 5)	(n = 2)	(n = 2)	(n = 3)	(n = 2)	(n = 6)	(n = 43)	< 0.001
Taking folic acid:										< 0.001
Prior to pregnancy	38.5	7.2	10.7	8.9	10.2	12.1	17.2	18.9	36.2
At confirmation	54.3	66.7	71.4	66.6	70.6	72.6	70.9	71.1	58.3
Not taking folic acid	7.1	26.1	17.8	24.5	19.2	15.4	11.9	9.9	5.5
Received infertility treatment										< 0.001
Length of the IPI:	3.7	1.2	0.4	1.7	1.3	2.4	1.6	2.3	3.3	< 0.001
< 12 months	17.2	29.4	21.1	24.5	20.4	16.7	13.7	13.2	17.8
12 to < 24 months	38.4	26.4	27.1	28.0	25.2	27.0	25.2	27.5	37.1
24 to < 36 months	23.8	15.9	20.2	19.6	21.4	20.0	19.6	23.8	23.1
36 months or more	20.6	28.2	31.6	28.0	32.9	36.4	41.5	35.4	22.0
Lone parent at P2	3.3	21.3	16.4	23.3	18.8	14.3	11.9	9.1	4.0	< 0.001
1^st infant birthweight, grams	3359.2	3161.7	3194.1	3180.9	3128.1	3312.7	3263.8	3418.6	3442.2	< 0.001
(mean, SD)	(524.0)	(552.2)	(554.4)	(492.6)	(492.8)	(516.1)	(551.3)	(530.3)	(538.2)	< 0.001
1^st infant SGA	12.0	22.5	20.6	22.5	19.8	16.4	14.9	8.6	9.4	< 0.001
1^st infant LGA	6.6	3.6	4.0	3.2	1.6	4.4	5.7	7.9	9.2	< 0.001
1^st infant PTB	4.9	5.7	6.4	5.8	6.4	5.5	6.4	4.2	5.0	0.253
2^nd infant birthweight, grams	3523.8	3214.4	3302.6	3226.1	3275.5	3364.9	3466.9	3557.8	3576.2	< 0.001
(mean, SD)	(511.2)	(544.6)	(535.8)	(534.7)	(505.0)	(530.5)	(551.9)	(538.8)	(512.0)	< 0.001
2^nd infant PTB	3.1	7.8	4.6	7.2	6.4	4.6	4.1	2.7	3.3	< 0.001
2^nd infant SGA	6.0	19.5	14.3	16.4	14.4	11.8	8.4	5.3	4.3	< 0.001
2^nd infant LGA	13.9	6.3	6.7	4.0	5.8	7.9	13.6	15.9	15.3	< 0.001

Open in a new tab

All figures are proportions (%), unless otherwise stated.

¹. A smoker at the first ANA for P2 who was not smoking at the first ANA for P1.

². A smoker at the first ANA for P1 who stopped before the first ANA for P2.

³. A smoker later in P1 or between pregnancies; not smoking at the first ANA for P1 or P2.

⁴. p-values calculated using ANOVA for continuous and Chi-squared tests for categorical variables.

Abbreviations: ANA, antenatal appointment; BMI, body mass index; IPI, interpregnancy interval (from P1 birth to P2 conception); IQR, inter-quartile range; LGA, large for gestational age (> 90^th percentile); P1, first pregnancy; P2, second pregnancy; PTB, preterm birth (< 259 days); SD, standard deviation; SGA, small for gestational age (< 10^th percentile).

Over 70% of women who reported smoking at the first antenatal appointment for their first pregnancy (n = 2522) also reported smoking at the first antenatal appointment for their second pregnancy (n = 1784). Those who smoked at the start of both their first two pregnancies accounted for 10.6% of all included women. A further 4.4% (n = 738) were categorised as smoker P1 (stopped before the first ANA P2) and 2.7% (n = 456) as smoker P2.

Mean maternal age at the start of the second pregnancy was the lowest for all categories of smokers (heavier smokers, (23.8 years, (standard deviation (SD) 4.4)), smokers (24.6 years (4.8)), smoker increased (23.4 years (3.9)) and smoker reduced (24.2 years (4.6)) and smoker P2 (24.7 years (4.8)). Mean maternal age at the second pregnancy was the highest in never smokers (30.2 years (5.0)) and ex-smokers (30.1 years (5.1)). At the start of the second pregnancy and compared to never smokers, all categories of smokers were more likely to be lone parents, of White ethnicity, of lower educational attainment, not to be taking folic acid in early pregnancy, and less likely to be in employment.

In terms of mothers’ ethnicity, our sample comprised 86.6% White, 5.9% Asian, 0.6% Chinese, 1.5% Black/African/Caribbean, and 1.2%. Mixed. Other ethnicities comprised 1.0% of the sample and 3.3% did not specify ethnicity.

The incidence of SGA birth for each of the first two pregnancies by maternal smoking status is shown in Fig 3 and in all cases, the prevalence is lower in the second pregnancy than in the first. The incidence of SGA birth in in never smokers was 12.0% in the first pregnancy and 6.0% in the second pregnancy. For ex-smokers these figures were 9.4% and 4.3% respectively. Of women who have never smoked and who had an SGA birth in the first pregnancy (n = 1004), over a quarter are of Asian ethnicity (n = 257). The incidence of first pregnancy SGA birth for the Asian women included in this study was 27.7%, compared to 11.2% for White women.

Table 3. Univariate odds of small for gestational age birth (< 10^th percentile) in the second pregnancy in the full sample, by maternal characteristics recorded at the start of or during the first and second pregnancies.

Maternal Characteristics	First pregnancy		Second pregnancy
Maternal Characteristics	n	OR (95% CI)	N	OR (95% CI)
Age category (ref = 25–34 years)
< 18 years	1005	2.2 (1.77, 2.73)	63	1.97 (0.89, 4.33)
18–24 years	5793	1.66 (1.45, 1.89)	3913	1.65 (1.44, 1.89)
35–39 years	845	0.99 (0.73, 1.36)	2386	1.01 (0.84, 1.22)
40 years and over	47	1.65 (0.59, 4.60)	326	0.81 (0.49, 1.35)
BMI category (ref = normal weight (18.5–24.9 kg/m ² ))
Underweight (<18.5 kg/m²)	628	1.93 (1.51, 2.46)	473	2.30 (1.77, 2.98)
Overweight (25–29.9 kg/m²)	4106	0.78 (0.67, 0.91)	4516	0.76 (0.65, 0.88)
Obese (≥30 kg/m²)	2241	0.71 (0.58, 0.87)	3282	0.69 (0.58, 0.82)
Highest level of education (ref = degree level)
College level	6272	1.20 (1.02, 1.41)	6923	1.31 (1.12, 1.52)
Secondary or below	5531	1.84 (1.57, 2.15)	4272	1.95 (1.66, 2.28)
Folic acid status (ref = taking prior to pregnancy)
Started once pregnancy confirmed	9624	1.55 (1.35, 1.79)	9986	1.47 (1.27, 1.70)
Not taking folic acid	1454	2.17 (1.76, 2.68)	1489	2.43 (1.99, 2.98)
Not in employment	3458	2.11 (1.85, 2.40)	5530	1.91 (1.69, 2.15)
Received infertility treatment	680	0.87 (0.63, 1.21)	514	0.75 (0.51, 1.11)
Lone parent	1450	1.42 (1.17, 1.72)	1057	1.44 (1.16, 1.78)
Gestational diabetes mellitus	292	0.43 (0.22, 0.84)	425	0.82 (0.54, 1.24)
Gestational hypertension	425	0.78 (0.51, 1.19)	188	1.74 (1.10, 2.74)
	Non pregnancy specific
	n	OR (95% CI)
Maternal Ethnicity (ref = White)
Mixed	196	0.98 (0.55, 1.77)
Asian	987	2.61 (2.16, 3.15)
Black/African/Caribbean	247	1.47 (0.94, 2.29)
Chinese	99	1.14 (0.53, 2.48)
Other	173	1.75 (1.07, 2.86)
Not known	551	0.93 (0.65, 1.33)
Length of the IPI (ref = 12 to < 24 months)
< 12 months	2937	1.32 (1.10, 1.57)
24 to < 36 months	3842	1.19 (1.01, 1.41)
36 months or more	4117	1.39 (1.19, 1.63)
Previous SGA birth	2067	6.67 (5.86, 7.58)

Open in a new tab

Abbreviations: BMI, body mass index; IPI, interpregnancy interval (from the birth of the first infant to the conception of the second); SGA, small for gestational age (< 10^th percentile); OR, odds ratio; CI, confidence interval.

Table 4 presents odds ratios for SGA birth in the second pregnancy according to the mother’s history of smoking and change in smoking behaviour between the first and second pregnancy, with Model 1 adjusting for confounders, and Model 2 adjusting for confounders and mediators.

Table 4. The odds of small for gestational age birth (< 10th percentile) in the second pregnancy.

	Full sample		Without previous SGA		With previous SGA
	n	Odds Ratios (95% CI)	n	Odds Ratios (95% CI)	n	Odds Ratios (95% CI)
Heavier Smoker	330	3.79 (2.84, 5.05)	256	3.66 (2.48, 5.41)	74	2.66 (1.64. 4.31)
Unadjusted		3.79 (2.84, 5.05)		3.66 (2.48, 5.41)		2.66 (1.64. 4.31)
Model 1^†		3.54 (2.55, 4.92)		3.53 (2.32, 5.38)		3.34 (1.96, 5.68)
Model 2^‡		3.57 (2.57, 4.97)		3.52 (2.31, 5.37)		3.54 (2.07, 6.08)
Smoker	786	2.64 (2.12, 3.29)	623	2.48 (1.84, 3.36)	163	1.93 (1.35, 2.74)
Unadjusted		2.64 (2.12, 3.29)		2.48 (1.84, 3.36)		1.93 (1.35, 2.74)
Model 1^†		2.44 (1.89, 3.15)		2.43 (1.75, 3.39)		2.34 (1.56, 3.51)
Model 2^‡		2.43 (1.88, 3.14)		2.47 (1.77, 3.44)		2.23 (1.48, 3.35)
Smoker increased	344	3.00 (2.21, 4.05)	267	2.88 (1.90, 4.37)	77	1.99 (1.22, 3.25)
Unadjusted		3.00 (2.21, 4.05)		2.88 (1.90, 4.37)		1.99 (1.22, 3.25)
Model 1^†		2.70 (1.92, 3.82)		2.84 (1.82, 4.44)		2.42 (1.41, 4.16)
Model 2^‡		2.75 (1.94, 3.88)		2.87 (1.84, 4.49)		2.51 (1.45, 4.32)
Smoker reduced	307	2.63 (1.89, 3.67)	247	3.02 (1.97, 4.61)	60	1.50 (0.84, 2.65)
Unadjusted		2.63 (1.89, 3.67)		3.02 (1.97, 4.61)		1.50 (0.84, 2.65)
Model 1^†		2.44 (1.68, 3.54)		2.98 (1.90, 4.67)		1.75 (0.94, 3.25)
Model 2^‡		2.50 (1.72, 3.63)		3.05 (1.95, 4.78)		1.82 (0.97, 3.39)
Smoker P2 ¹	452	2.09 (1.55, 2.82)	377	2.22 (1.50, 3.28)	75	1.54 (0.92, 2.58)
Unadjusted		2.09 (1.55, 2.82)		2.22 (1.50, 3.28)		1.54 (0.92, 2.58)
Model 1^†		2.11 (1.51, 2.95)		2.22 (1.47, 3.37)		1.93 (1.11, 3.36)
Model 2^‡		2.13 (1.52, 2.98)		2.26 (1.49, 3.42)		1.91 (1.09, 3.34)
Smoker P1 ²	735	1.45 (1.10, 1.91)	626	1.75 (1.24, 2.46)	109	0.88 (0.54, 1.44)
Unadjusted		1.45 (1.10, 1.91)		1.75 (1.24, 2.46)		0.88 (0.54, 1.44)
Model 1^†		1.50 (1.10, 2.03)		1.75 (1.22, 2.53)		1.05 (0.62, 1.78)
Model 2^‡		1.53 (1.13, 2.07)		1.80 (1.25, 2.60)		1.05 (0.62, 1.79)
Other smoker ³	1346	0.89 (0.69, 1.15)	1230	1.09 (0.80, 1.48)	116	0.82 (0.50, 1.33)
Unadjusted		0.89 (0.69, 1.15)		1.09 (0.80, 1.48)		0.82 (0.50, 1.33)
Model 1^†		1.11 (0.84, 1.45)		1.17 (0.84, 1.62)		0.98 (0.59, 1.64)
Model 2^‡		1.12 (0.85, 1.47)		1.17 (0.85, 1.62)		1.01 (0.61, 1.69)
Ex-smoker	4065	0.70 (0.59, 0.84)	3681	0.81 (0.65, 1.01)	384	0.65 (0.47, 0.88)
Unadjusted		0.70 (0.59, 0.84)		0.81 (0.65, 1.01)		0.65 (0.47, 0.88)
Model 1^†		0.89 (0.73, 1.07)		0.93 (0.74, 1.17)		0.82 (0.59, 1.15)
Model 2^‡		0.90 (0.74, 1.08)		0.93 (0.74, 1.17)		0.83 (0.60, 1.17)
Never smoker	8354	Reference	7353	Reference	1001	Reference

Open in a new tab

¹. A smoker at the first ANA for P2 who was not smoking at the first ANA for P1.

². A smoker at the first ANA for P1 who stopped before the first ANA for P2.

³. A smoker later in P1 or between pregnancies; not smoking at the first ANA for P1 or P2.

^† Model 1 (adjusts for confounders): Adjusted for maternal age, BMI, educational attainment, employment status, partnership status, folate supplementation and infertility treatment at the start of the first pregnancy, gestational diabetes mellitus and gestational hypertension recorded during the first pregnancy, SGA birth in the first pregnancy (not in the stratified analysis), maternal ethnicity and the length of the interpregnancy interval.

^‡ Model 2 (adjusts for confounders and mediators): Adjusted for maternal age, BMI, educational attainment, employment status, partnership status, folate supplementation and infertility treatment at the start of the first pregnancy, gestational diabetes mellitus and gestational hypertension recorded during the second pregnancy, SGA birth in the first pregnancy (not in the stratified analysis), maternal BMI at the start of the second pregnancy, maternal ethnicity and the length of the interpregnancy interval.

Abbreviations: ANA, antenatal appointment; BMI, body mass index; CI, confidence interval; P1, first pregnancy; P2 second pregnancy; SGA, small for gestational age (<10^th percentile).

Model 1 adjusts for confounders and in the full sample shows the odds of SGA birth in the second pregnancy adjusting for maternal age, BMI, educational attainment, employment status, partnership status, folate supplementation and infertility treatment at the start of the first pregnancy, gestational diabetes mellitus and gestational hypertension recorded during the first pregnancy, SGA birth in the first pregnancy, maternal ethnicity and the length of the interpregnancy interval.

Compared to never smokers, there are increased odds of SGA birth in the second pregnancy for heavier smokers ((adjusted odds ratio (aOR) 3.54 [95% confidence interval (CI) 2.55, 4.92]), smokers (2.44 [1.89, 3.15]), smoker increased (2.70 [1.92, 3.82]), smoker reduced (2.44 [1.68, 3.54]), smokers P2 (2.11 [1.51, 2.95]) and smokers P1 (stopped before the first ANA P2) (1.50 [1.10, 2.03]). Other smokers, (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) or ex-smokers did not have increased odds of SGA birth in the second pregnancy compared to never smokers ((1.11 [0.84, 1.45]) and (0.89 [0.73, 1.07]) respectively).

Model 1 in the sample which excludes women whose first child was born SGA makes the same adjustments described above, with the exception of an adjustment for previous SGA birth. Compared to never smokers, there were increased odds of new SGA for heavier smokers (3.53 [2.32, 5.38]), smokers (2.43 [1.75, 3.39]), smoker increased (2.84 [1.82, 4.44]), smoker reduced (2.98 [1.90, 4.67]), smokers P2 (2.22 [1.47, 3.37]) and smokers P1 (stopped before the first ANA P2) (1.75 [1.22, 2.53]). Other smokers (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) or ex-smokers did not have increased odds of new SGA compared to never smokers ((1.17 [0.84, 1.62]) and (0.93 [0.74, 1.17]) respectively).

Model 1 in the sample where there was SGA birth in the first pregnancy, shows the odds of recurrent SGA birth. The same adjustments were made as described above.

Compared to never smokers, there were increased odds of recurrent SGA birth in the second pregnancy heavier smokers (3.34 [1.96, 5.68]), smokers (2.34 [1.56, 3.51]), smoker increased (2.42 [1.41, 4.16]) and smokers P2 (1.93 [1.11, 3.36]). Compared to never smokers, there was no increase in the odds of recurrent SGA birth for smoker reduced (1.75 [0.94, 3.25]), smokers P1 (stopped before the first ANA P2) (1.05 [0.62, 1.78]), other smokers (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) (0.98 [0.59, 1.64]) or ex-smokers (0.82 [0.59, 1.15]).

Model 2 adjusts for confounders and mediators and in the full sample shows the odds of SGA birth in the second pregnancy adjusting for maternal age, BMI, educational attainment, employment status, partnership status, folate supplementation and infertility treatment at the start of the first pregnancy, gestational diabetes mellitus and gestational hypertension recorded during the second pregnancy, maternal BMI at the start of the second pregnancy, SGA birth in the first pregnancy, maternal ethnicity and the length of the interpregnancy interval.

Compared to never smokers, there were increased odds of SGA birth in the second pregnancy for heavier smokers (3.57 [2.57, 4.97]), smokers (2.43 [1.88, 3.14]), smoker increased (2.75 [1.94, 3.88]), smoker reduced (2.50 [1.72, 3.63]), smokers P2 (2.13 [1.52, 2.98]) and smokers P1 (stopped before the first ANA P2) (1.53 [1.13, 2.07]). Other smokers, (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) or ex-smokers did not have increased odds of SGA birth in the second pregnancy compared to never smokers ((1.12 [0.85, 1.47]) and (0.90 [0.74, 1.08]) respectively).

Model 2 in the sample which excluding women whose first child was born SGA makes the same adjustments described above, with the exception of an adjustment for previous SGA birth. Compared to never smokers, there were increased odds of new SGA for heavier smokers (3.52 [2.31, 5.37]), smokers (2.47 [1.77, 3.44]), smoker increased (2.87 [1.84, 4.49]), smoker reduced (3.05 [1.95, 4.78]), smokers P2 (2.26 [1.49, 3.42]) and smokers P1 (stopped before the first ANA P2) (1.80 [1.25, 2.60]). Other smokers (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) or ex-smokers did not have increased odds of new SGA compared to never smokers ((1.17 [0.85, 1.62]) and (0.93 [0.74, 1.17]) respectively).

Model 2 in the sample with SGA birth in the first pregnancy shows the odds of recurrent SGA birth. The same adjustments were made as described above.

Compared to never smokers, there were increased odds of recurrent SGA birth in the second pregnancy heavier smokers (3.54 [2.07, 6.08]), smokers (2.23 [1.48, 3.35]), smoker increased (2.51 [1.45, 4.32]) and smokers P2 (1.91 [1.09, 3.34]). Compared to never smokers, there was no increase in the odds of recurrent SGA birth for smoker reduced (1.82 [0.97, 3.39]), smokers P1 (stopped before the first ANA P2) (1.05 [0.62, 1.79]), other smokers (smokers later in P1 or between pregnancies but not smoking at the first ANA of P1 or P2) (1.01 [0.61, 1.69]) or ex-smokers (0.83 [0.60, 1.17]).

The full results for Model 1 in the full sample (Table 4) are given in Table 5.

Table 5. Full results of Model 1 in Table 4; the adjusted odds of small for gestational age birth (<10^th percentile) in the second pregnancy in the full sample.

	aOR	95% CI
Maternal smoking status (ref = never smoked)
Heavier smoker	3.54	2.55	4.92
Smoker	2.44	1.89	3.15
Smoker increased	2.70	1.92	3.82
Smoker reduced	2.44	1.68	3.54
Smoker P2¹	2.11	1.51	2.95
Smoker P1²	1.50	1.10	2.03
Other smoker³	1.11	0.84	1.45
Ex-smoker	0.89	0.73	1.07
Maternal age at booking	1.00	0.99	1.02
Maternal BMI	0.98	0.96	0.99
In employment	0.83	0.70	0.97
Lone parent	0.93	0.75	1.16
Previous SGA birth	5.48	4.79	6.26
Gestational Diabetes	0.42	0.20	0.86
Gestational Hypertension	0.83	0.53	1.29
Received infertility treatment	0.88	0.62	1.26
Length of the IPI (days)	1.00	1.00	1.00
Maternal ethnicity (ref = White)
Mixed	0.95	0.52	1.75
Asian	2.09	1.66	2.63
Black/African/Caribbean	1.47	0.92	2.36
Chinese	1.37	0.62	3.05
Other	1.86	1.11	3.15
Not known	1.01	0.69	1.47
Maternal education (ref = Degree)
College level	0.97	0.81	1.17
Secondary or below	1.06	0.87	1.29
Folic acid (ref = taking prior to pregnancy)
Started once pregnancy confirmed	1.10	0.93	1.29
Not taking folic acid	1.19	0.93	1.52

Open in a new tab

¹. A smoker at the first ANA for P2 who was not smoking at the first ANA for P1.

². A smoker at the first ANA for P1 who stopped before the first ANA for P2.

³. A smoker later in P1 or between pregnancies; not smoking at the first ANA for P1 or P2.

Abbreviations: ANA, antenatal appointment; BMI, body mass index; IPI, interpregnancy interval (from the birth of the first infant to the conception of the second); P1, first pregnancy; P2, second pregnancy; SGA, small for gestational age (< 10^th percentile); aOR, adjusted odds ratio; CI, confidence interval.

Sensitivity analysis for Model 1 was run using the other minimal adjustment sets identified by DAGitty as described in the Methods section above [22]. The results of this analysis (S1 Table) show only very minor differences in the adjusted odds ratios for Model 1 whichever minimal adjustment set is used noting slight differences in the numbers of missing observations across the different models.

Discussion

In the overall sample we found that mothers smoking at the start of the first pregnancy had a 50% higher risk of SGA birth in the second pregnancy compared to never smokers even if the mother stopped smoking before the first antenatal appointment of the second pregnancy. However, if the mother was not a smoker at the first antenatal appointment for either her first or her second pregnancy, but smoked either later in her first pregnancy or between pregnancies, there was no evidence of increased risk of SGA in her second pregnancy compared to never smokers. When we stratified by previous SGA, this was true for new SGA birth but not for recurrent SGA birth.

According to this analysis, smoking at the start of the first pregnancy may be an important factor in shaping the risk of SGA birth in the second pregnancy. It should be noted, however, that mothers smoking at the start of their first pregnancies could have quit smoking at any point after the first antenatal appointment for their first pregnancy, right up until they found out that they were pregnant for the second time (Table 1).

In all the analyses, second infants born to mothers who reported smoking at the start of both of their first two pregnancies were more likely to be born SGA compared to those of never smokers, with the highest odds of SGA birth found for the heaviest smokers at the start of both pregnancies.

In the analysis of recurrent SGA birth, smokers who reported smoking fewer cigarettes a day at the start of their second pregnancy than they did at the start of their first pregnancy, or who smoked at the start of their first pregnancy but quit by the latest when the second pregnancy was confirmed did not have increased odds of a second infant being born SGA. We do not know however whether these women will have actually quit smoking at some later stage during pregnancy to help avoid a recurrent SGA birth.

Maternal smoking is self-reported and there may be an element of under-reporting. Women could either still be smoking at the start of their second pregnancies or resume later during the pregnancy. A comparison of concurrent and retrospective self-reports of smoking status in pregnancy found 19% of all discordant reports (total n = 222) were where mothers recalled smoking daily in pregnancy but had not reported this at the time of their pregnancy and an additional 39% reported occasional smoking where they had registered as non-smokers in pregnancy [35]. The remaining discordant reports were where mothers failed to recall smoking which they had reported in pregnancy [35]. The study found that younger mothers, multiparae, those with lower levels of educational attainment and those who were not in a stable relationship had lower concordance on reports of smoking in pregnancy compared to older mothers, primiparae, those who were more highly educated and those living with the father at the time of pregnancy respectively [35]. In our study women were asked for their smoking status at the start of each pregnancy and the responses recorded at that time, which means that recall bias is unlikely.

We found similar a similar percentage of women smoked at the start of both of their first two pregnancies to those reported elsewhere [36, 37]. Whilst the time between pregnancies, where a women is still in relatively intense contact with healthcare professionals, is the ideal time to focus on the health of the entire family, particularly for mothers with a previous history of SGA birth, this is obviously a missed opportunity. Whilst mothers who were smoking at the start of their first pregnancy still have an increased risk of SGA birth in their second pregnancy the risk is lower than for those continuing to smoke at the start of the second pregnancy.

Healthcare professionals can refer pregnant smokers to smoking cessation services but there are a number of areas which could be considered and evaluated further. These include smoking support for entire family groups [38]. Financial incentives and rewards have been shown to have a positive impact on increasing long-term rates of smoking cessation in pregnancy and the post-partum period [39]. The use of financial and other interventions, including social media applications, websites and text messaging, have received mixed feedback depending upon whether this was sought from mothers, significant others (including partners) or healthcare professionals [38]. Targeted leaflets, posters and campaigns could be a useful persuasive tool particularly where the specific effects of smoking on the developing fetus are emphasized [38].

Strengths and limitations

Our study has a number of strengths. The SLOPE study is a large population-based cohort which includes women from all socio-economic and ethnic backgrounds which is representative of the regional population. The ethnic make-up of our sample is comparable with the 2011 England and Wales census with 86% White, 7.5% Asian/Asian British (which includes Chinese), 3.3% Black/African/Caribbean/Black British and 2.2% Mixed/multiple ethnic groups [40].

The Southampton data observatory reports that, based on the 2019 Indices of Deprivation published by the Ministry of Housing Communities and Local Government [41], Southampton is currently ranked 55^th out of 317 local authorities based on the average neighbourhood deprivation rank and approximately 45% of the Southampton’s population reside in areas which fall within the 30% most deprived nationally [42]. In this analysis approximately half of the women live in Southampton, with half living in the rest of Hampshire which is less deprived.

The analysis was able to adjust for several key confounders and outcome measurements were based on records which were objectively measured by healthcare professionals.

There are some limitations, primarily the fact that the majority of variables were self-reported. Using self-reported maternal smoking status in pregnancy means that there is the possibility of non-disclosure and information bias affecting the ability to characterise the exposure correctly [43]. Suggested methods of overcoming these potential biases are also subject to a number of issues. For example, biologic assays are considered a more accurate way of measuring maternal smoking but still only reflect exposure over short periods and variations in nicotine metabolism affect the net exposure [43].

We were also unable to incorporate risk factors for smoking continuation, inception and cessation such as having a partner who smokes (potentially a different partner to the first pregnancy), other smoking within the household and other exposure to passive smoking.

Repeating this analysis in other datasets will enable the comparison of results to see if our findings are replicated elsewhere.

Conclusion

In the analysis of the full sample and in women without a previous SGA birth, smoking in the first pregnancy was associated with increased odds of having a SGA infant in the second pregnancy, even if the mother did not report smoking at the first antenatal appointment of the second pregnancy. Where a mother quit smoking at any point up to the confirmation of the second pregnancy, the odds were lower than for women continuing to smoke or those who smoked at the start of their second pregnancy only (compared to never smokers).

In women who were smokers in their first pregnancy and who gave birth to their first infant who was SGA, there was no increase in the odds of having a further SGA infant in the second pregnancy where they quit smoking at any point up to the confirmation of the second pregnancy or where the number of cigarettes a day was reduced from 10 or more in the first pregnancy to up to 10 a day in the second pregnancy (compared to never smokers).

Interventions which support mothers to stop smoking between pregnancies or at the start of her second pregnancy or which help her to reduce the number of cigarettes smoked a day may help to reduce the incidence of having a SGA infant in the second pregnancy.

Supporting information

S1 Table. Sensitivity analysis showing the effect of using different minimal adjustment sets on the adjusted odds ratios calculated in Model 1 in the full sample.

(DOCX)

Click here for additional data file.^{(24.1KB, docx)}

Acknowledgments

We thank David Cable (Electronic Patient Records Implementation and Service Manager) and Florina Borca (Senior Information Analyst R&D) at University Hospital Southampton NHS Foundation Trust for support in accessing the data used in this study.

Abbreviations

ANA: Antenatal appointment
BMI: Body mass index
DAG: Directed Acyclic Graph
GDM: Gestational diabetes mellitus
P1: First pregnancy
P2: Second pregnancy
SGA: Small for gestational age (< 10^th percentile)
SLOPE: Studying Lifecourse Obesity PrEdictors

Data Availability

The study’s ethical approval from the Faculty of Medicine Ethics Committee, University of Southampton and the Health Research Authority restricts public sharing of the raw data used in this study. To request access conditional on approval from the appropriate institutional ethics, research governance processes and data owners, please email rgoinfo@soton.ac.uk.

Funding Statement

This research is supported by an Academy of Medical Sciences and Wellcome Trust grant to NAA [Grant no: AMS_HOP001\1060] (https://acmedsci.ac.uk/) (https://wellcome.ac.uk) and an NIHR Southampton Biomedical Research Centre PhD studentship to EJT. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Cnattingius S. The epidemiology of smoking during pregnancy: Smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine & Tobacco Research. 2004;6(Suppl_2):S125–S40. [DOI] [PubMed] [Google Scholar]
2.Zdravkovic T, Genbacev O, McMaster MT, Fisher SJ. The adverse effects of maternal smoking on the human placenta: A review. Placenta. 2005;26:S81–S6. doi: 10.1016/j.placenta.2005.02.003 [DOI] [PubMed] [Google Scholar]
3.Kolås T, Nakling J, Salvesen KA. Smoking during pregnancy increases the risk of preterm births among parous women. Acta Obstet Gynecol Scand. 2000;79(8):644–8. [PubMed] [Google Scholar]
4.Ananth CV, Smulian JC, Demissie K, Vintzileos AM, Knuppel RA. Placental abruption among singleton and twin births in the United States: risk factor profiles. Am J Epidemiol. 2001;153(8):771–8. doi: 10.1093/aje/153.8.771 [DOI] [PubMed] [Google Scholar]
5.Kyrklund-Blomberg NB, Cnattingius S. Preterm birth and maternal smoking: Risks related to gestational age and onset of delivery. American Journal of Obstetrics & Gynecology. 1998;179(4):1051–5. doi: 10.1016/s0002-9378(98)70214-5 [DOI] [PubMed] [Google Scholar]
6.Mitchell EA, Thompson JMD, Robinson E, Wild CJ, Becroft DMO, Clark PM, et al. Smoking, nicotine and tar and risk of small for gestational age babies. Acta Paediatrica. 2002;91(3):323–8. doi: 10.1080/08035250252834003 [DOI] [PubMed] [Google Scholar]
7.Lie RT, Wilcox AJ, Taylor J, Gjessing HK, Saugstad OD, Aabyholm F, et al. Maternal smoking and oral clefts: the role of detoxification pathway genes. Epidemiology (Cambridge, Mass). 2008;19(4):606–15. doi: 10.1097/EDE.0b013e3181690731 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Lange S, Probst C, Rehm J, Popova S. National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis. The Lancet Global Health. 2018;6(7):e769–e76. doi: 10.1016/S2214-109X(18)30223-7 [DOI] [PubMed] [Google Scholar]
9.NHS Digital. Statistics on Women’s Smoking Status at Time of Delivery: England Quarter 3, 2019–20 [accessed 24 March 2020]. https://digital.nhs.uk/data-and-information/publications/statistical/statistics-on-women-s-smoking-status-at-time-of-delivery-england/statistics-on-womens-smoking-status-at-time-of-delivery-england-quarter-3-2019-20.
10.Rousseaux S, Seyve E, Chuffart F, Bourova-Flin E, Benmerad M, Charles MA, et al. Immediate and durable effects of maternal tobacco consumption alter placental DNA methylation in enhancer and imprinted gene-containing regions. BMC medicine. 2020;18(1):306. doi: 10.1186/s12916-020-01736-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Nakamura A, Francois O, Lepeule J. Epigenetic Alterations of Maternal Tobacco Smoking during Pregnancy: A Narrative Review. Int J Environ Res Public Health. 2021;18(10). doi: 10.3390/ijerph18105083 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Kvalvik LG, Haug K, Klungsøyr K, Morken N-H, DeRoo LA, Skjærven R. Maternal Smoking Status in Successive Pregnancies and Risk of Having a Small for Gestational Age Infant. Paediatric and Perinatal Epidemiology. 2017;31(1):21–8. doi: 10.1111/ppe.12333 [DOI] [PubMed] [Google Scholar]
13.Fleming TP, Watkins AJ, Velazquez MA, Mathers JC, Prentice AM, Stephenson J, et al. Origins of lifetime health around the time of conception: causes and consequences. The Lancet. 2018;391(10132):1842–52. doi: 10.1016/S0140-6736(18)30312-X [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Mathews F, Yudkin P, Smith RF, Neil A. Nutrient intakes during pregnancy: the influence of smoking status and age. Journal of Epidemiology and Community Health. 2000;54(1):17. doi: 10.1136/jech.54.1.17 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Ziauddeen N, Wilding S, Roderick PJ, Macklon NS, Smith D, Chase D, et al. Predicting the risk of childhood overweight and obesity at 4–5 years using population-level pregnancy and early-life healthcare data. BMC medicine. 2020;18(1):105. doi: 10.1186/s12916-020-01568-z [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Wilding S, Ziauddeen N, Smith D, Roderick P, Chase D, Alwan NA. Are environmental area characteristics at birth associated with overweight and obesity in school-aged children? Findings from the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the south of England. BMC medicine. 2020;18(1):43. doi: 10.1186/s12916-020-01513-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Wilding S, Ziauddeen N, Roderick P, Smith D, Chase D, Macklon N, et al. Are socioeconomic inequalities in the incidence of small-for-gestational-age birth narrowing? Findings from a population-based cohort in the South of England. BMJ Open. 2019;9(7):e026998. doi: 10.1136/bmjopen-2018-026998 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.National Institute for Care and Clinical Excellence. Antenatal care for uncomplicated pregnancies 2019 [accessed 30 June 2021]. https://pathways.nice.org.uk/pathways/antenatal-care-for-uncomplicated-pregnancies. [PubMed]
19.Norris T, Seaton SE, Manktelow BN, Baker PN, Kurinczuk JJ, Field D, et al. Updated birth weight centiles for England and Wales. Arch Dis Child Fetal Neonatal Ed. 2018;103(6):F577–F82. doi: 10.1136/archdischild-2017-313452 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.World Health Organisation: Report of a WHO technical consultation on birth spacing: Geneva, Switzerland 13–15 June W2005. World Health Organization.: World Health Organization; 2007.
21.Textor J, van der Zander B, Gilthorpe MS, Liskiewicz M, Ellison GT. Robust causal inference using directed acyclic graphs: the R package ’dagitty’. Int J Epidemiol. 2016;45(6):1887–94. doi: 10.1093/ije/dyw341 [DOI] [PubMed] [Google Scholar]
22.Textor J. DAGitty—draw and analyze causal diagrams [Website]. http://www.dagitty.net/.
23.Suttorp MM, Siegerink B, Jager KJ, Zoccali C, Dekker FW. Graphical presentation of confounding in directed acyclic graphs. Nephrology Dialysis Transplantation. 2015;30(9):1418–23. doi: 10.1093/ndt/gfu325 [DOI] [PubMed] [Google Scholar]
24.Textor J. Drawing and Analyzing Causal DAGs with DAGitty [updated 11 January 2020. Manual]. http://dagitty.net/manual-3.x.pdf.
25.McCowan L, Horgan RP. Risk factors for small for gestational age infants. Best Practice & Research Clinical Obstetrics & Gynaecology. 2009;23(6):779–93. [DOI] [PubMed] [Google Scholar]
26.R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria: 2020. [Google Scholar]
27.Dowle M, Srinivasan A. data.table: Extension of ‘data.frame’. R package version 1.13.0. 2020. [Google Scholar]
28.Wickham H, François R, Henry L, Müller K. dplyr: A Grammar of Data Manipulation. R package version 1.0.2. 2020. [Google Scholar]
29.Chongsuvivatwong V. epiDisplay: Epidemiological Data Display Package. R package version 3.5.0.1. 2018. [Google Scholar]
30.Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag; New York. 2016. [Google Scholar]
31.Wickham H, Miller E. haven: Import and Export ’SPSS’, Stata’ and ’SAS’ Files. 2020.
32.Revelle W. psych: Procedures for Personality and Psychological Research, Northwestern University, Evanston, Illinois, USA. [Google Scholar]
33.Wickham H. Reshaping data with the reshape package. Journal of Statistical Software, 21(12). 2007. [Google Scholar]
34.Wickham H. tidyr: Tidy Messy Data. R package version 1.1.2. 2020. [Google Scholar]
35.Post A, Gilljam H, Bremberg S, Galanti MR. Maternal smoking during pregnancy: A comparison between concurrent and retrospective self-reports. Paediatric and Perinatal Epidemiology. 2008;22(2):155–61. doi: 10.1111/j.1365-3016.2007.00917.x [DOI] [PubMed] [Google Scholar]
36.Tran DT, Roberts CL, Jorm LR, Seeho S, Havard A. Change in smoking status during two consecutive pregnancies: a population-based cohort study. BJOG: An International Journal of Obstetrics & Gynaecology. 2014;121(13):1611–20. doi: 10.1111/1471-0528.12769 [DOI] [PubMed] [Google Scholar]
37.Hauge LJ, Aarø LE, Torgersen L, Vollrath ME. Smoking during consecutive pregnancies among primiparous women in the population-based Norwegian Mother and Child Cohort Study. Nicotine Tob Res. 2013;15(2):428–34. doi: 10.1093/ntr/nts147 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Bauld L, Graham H, Sinclair L, Flemming K, Naughton F, Ford A, et al. Barriers to and facilitators of smoking cessation in pregnancy and following childbirth: literature review and qualitative study. Health Technol Assess. 2017;21(36):1–158. doi: 10.3310/hta21360 [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Notley C, Gentry S, Livingstone-Banks J, Bauld L, Perera R, Hartmann-Boyce J. Incentives for smoking cessation. Cochrane Database of Systematic Reviews. 2019(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Office for National Statistics. Ethnicity and National Identity in England and Wales: 2011 [accessed 24 March 2020]. https://www.ons.gov.uk/peoplepopulationandcommunity/culturalidentity/ethnicity/articles/ethnicityandnationalidentityinenglandandwales/2012-12-11.
41.English indices of deprivation 2019 [accessed 1 September 2020]. https://www.gov.uk/government/statistics/english-indices-of-deprivation-2019.
42.Southampton data observatory [accessed 1 September 2020]. https://data.southampton.gov.uk/economy/deprivation-poverty/.
43.Dukic VM, Niessner M, Pickett KE, Benowitz NL, Wakschlag LS. Calibrating self-reported measures of maternal smoking in pregnancy via bioassays using a Monte Carlo approach. Int J Environ Res Public Health. 2009;6(6):1744–59. doi: 10.3390/ijerph6061744 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0260134.r001

Decision Letter 0

JJ Cray Jr

15 Jul 2021

PONE-D-21-13251

Maternal smoking behaviour across the first two pregnancies and small for gestational age birth: analysis of the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the South of England

PLOS ONE

Dear Dr. Taylor,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Enclosed you will find thorough reviews from two experts in the field. They identified major concerns with the manuscript as submitted. Both reviewers and I believe that if the authors are too address the raised points, constituting a major revision, this work would likely meet Plos One's publication criterion. Note major themes of concern were apriori vs post-hoc hypothesis or hypothesis generating approaches, confounding nature of patient self report data, concerns over replicability of the models, co-variate inclusions, and lack of descriptive data.

Please submit your revised manuscript by Aug 29 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

JJ Cray Jr., Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

If you are reporting a retrospective study of medical records or archived samples, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

3. Thank you for stating the following in the Competing Interests section:

[I have read the journal's policy and the authors of this manuscript have the following competing interests: KMG has received reimbursement for speaking at conferences sponsored by companies selling nutritional products and is part of an academic consortium that has received research funding from BenevolentAI Bio Ltd, Abbott Nutrition, Nestec and Danone. The other authors have no potentially competing interests to declare.].

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

4. We noted in your submission details that a portion of your manuscript may have been presented or published elsewhere.

[This manuscript is not currently being considered for publication by any other journal but an Abstract based on this analysis was accepted by the Society for Social Medicine and Population Health for their 2020 Annual Scientific Conference and was published in the Journal of Epidemiology and Community Health (Taylor E, Ziauddeen N, Godfrey K, et al OP64 Change in maternal smoking behaviour between two pregnancies and small for gestational age birth: analysis of a UK population-based cohort J Epidemiol Community Health 2020;74:A30-A31.).]

Please clarify whether this conference proceeding was peer-reviewed and formally published. If this work was previously peer-reviewed and published, in the cover letter please provide the reason that this work does not constitute dual publication and should be included in the current manuscript.

5. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

6. We note that the grant information you provided in the ‘Funding Information’ and ‘Financial Disclosure’ sections do not match.

When you resubmit, please ensure that you provide the correct grant numbers for the awards you received for your study in the ‘Funding Information’ section.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Introduction:

This is an interesting retrospective study from a large prospectively collected dataset. In my opinion its value is as an exploratory study for hypothesis generation. Intuitively, I would have expected that women ceasing smoking between their first and second pregnancies would reduce their risk of SGA in the second pregnancy to around those of ex-smokers or never smokers. Instead, the authors have hypothesised that they have a continued increased risk for SGA when compared to never smokers without clarifying or justifying their hypothesis.

The fact they have shown an increased risk is hypothesis generating but I feel the conclusion is tentative and needs repeating in other data sets. There is significant potential for the result to be biased, based on patient recall of smoking status and the complicated stratification of smoking status that the authors have proposed.

I am also challenged by their decisions regarding covariates on which to adjust the odds ratios. The DAG diagram they present is overly complex and the software used to create a parsimonious set of adjusting covariates is not explained adequately. Their approach here negates any advantage of the DAG process to create a discussion and to explicitly explain and argue for the choices they have made. I would not feel comfortable with the current models without significant additional explanation of their decisions to adjust on covariates present at the start of the first pregnancy not the pregnancy of interest.

The hypothesis:

The stated hypothesis…. “that mothers who smoked in a previous pregnancy or who smoked between pregnancies have a higher risk of SGA in the second pregnancy compared to never smokers, even if they were not smoking during the second pregnancy”. Seems not to be an a priori hypothesis. It seems rather to have come out of the authors analysis of the data. The literature that the authors cite does not support this hypothesis.

I do agree with the authors stated aim… “to characterise maternal smoking behaviours across a mother’s first two pregnancies and examine the relation of smoking behaviours with second child’s risk of being born SGA”.

In my opinion the study would be better characterised as exploratory and hypothesis generating.

The 11 smoking categories for the Exposure are too complicated:

The following are clear and well defined:

• Heavy smoker

• Smoker

• Smoker increased

• Smoker reduced

• Smoker P2

• Ex smoker

• Never smoker

The following are likely to have significant overlap give that the classifications are self reported as well as being subject to recall bias.

• Smoker P1

• Serial quitter

• Smoker later in P1 or between pregnancies

I would like to see the impact of collapsing these into a single category of Smoked between pregnancies; quit prior to or on confirmation P2.

Exploratory analysis of collapsing these groups shows an unadjusted odds ratio for SGA when compared to the never smoked group of OR=1.08; 95% CI [.88-1.31].

As the aim is to examine the relationship between smoking behaviours between pregnancies and the outcomes for P2 whether the following behaviours could also be collapsed to simply things. Smoker – Heavy smoker + Smoker; Not smoking – Ex smoker+ Never smoker. I appreciate that there is a dose-response gradient for the smokers but as the studies stated aim is to examine changes across pregnancies, the inclusion of 11 categories makes the paper complex and more challenging to follow and a more parsimonious approach may help readers.

The use of Directed Acyclic Graph (DAG) technique:

I appreciate that the DAG technique allows for discussion by making underlying relations explicit. However, Figure 2 is incredibly complex and difficult to follow. Further, the use of the software package “DAGitty” to create a parsimonious set of factors to adjust for is poorly explained and has the feeling of a “black box” approach that is opaque to the reader.

At face value I struggle with the author’s or DAGitty’s choice to use variables (Age, Education, Infertility treatment, folic acid) from P1 for the adjusted odds ratio. I struggle even more with the use of complications in the first pregnancy (Gestational diabetes mellitus, Gestational hypertension).

I would appreciate a more explicit explanation of this decision and greater explanation of the DAGitty process and the intervening steps. A more parsimonious diagram of the final proposed causal model would help the reader decide whether they feel the approach and hence any results and conclusions drawn is appropriate.

Table 3:

Could the maternal characteristics at the start of pregnancy 2 be expressed as univariate odds ratios as per Table 6?

This would usually allow the reader to examine whether there is a statistical basis for adjustment on the covariates present at the beginning of the second pregnancy. Notwithstanding that there may be other reasons for inclusion of a covariate based on content knowledge or prior research.

Why did the authors present this table allowing the reader to examine the relationship between maternal characteristics at the start of the second pregnancy when these are not the covariates that they use to calculate the adjusted odds ratios?

Table 5:

In my opinion this sensitivity analysis adds nothing to the study. I would omit it or move it to a supplementary analysis.

Discussion:

As previously discussed, I do not think that there is enough a priori evidence for the authors main hypothesis. Therefore, I feel their conclusion at the beginning of the discussion section is too strong.

This paper is a good exploration of their data set but given the nature of the smoking status variable a sensitivity analysis for all variations of the smoking status that includes any smoking between pregnancies would lead to a null finding.

The authors own discussion acknowledges the issues with patient recall of smoking around pregnancy with discordance rates between 19-39%. Further, it seems that women with higher risk factors for SGA were more likely to have discordance between reported and actual rates of smoking in pregnancy.

Conclusion:

I would recommend a major revision and further explanation of the multivariate models before publication. I would be happy to review a revised draft of this paper.

Reviewer #2: This is a well conducted and described study. I have a few minor comments:

- the counts of individuals in each of the exposure categories - included in the description of the exposure variable in the first column of Table 1 - are results and should not be included in the methods section. They are already included in the first row of table 2 in the results section, so it would be sufficient to just remove them from table 1.

- please remove the ± designation wherever you report SDs and instead report the means followed by the SDs in brackets.

- in Table 2 where you report proportions (e.g. of BMI, ethnicity, education categories), it is unclear what the numbers included in brackets represent. Additionally for the last row where you report n's again, it would be better to instead report the numbers of missing, since the n's are already included in the column headers.

- because the results in Table 2 are purely descriptive (not inferential), you should not be reporting any confidence intervals here. You should simply report mean (SD) or median (IQR) for continuous variables and n (%) for categorical ones (where n is the count of each category and % is the percentage of that count out of the number of individuals in the main exposure/column)

- relating to the comment above, please distinguish clearly between N, the denominator for a proportion, and n the numerator; so for example, you shouldn't have two columns both labelled n as you do in table 3. The first n is the denominator and should be labelled capital N.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Adam Mackie

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Nov 18;16(11):e0260134. doi: 10.1371/journal.pone.0260134.r002

Author response to Decision Letter 0

8 Oct 2021

Please see the uploaded response to reviewer comments document and also the cover letter dated 8 October 2021. Please do not hesitate to contact me if I can provide any futher details.

Best wishes

Elizabeth Taylor

E.J.Taylor@soton.ac.uk

Attachment

Submitted filename: Response to reviewers 2021_09_14.docx

Click here for additional data file.^{(942.1KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0260134.r003

Decision Letter 1

JJ Cray Jr

4 Nov 2021

PONE-D-21-13251R1

Dear Dr. Taylor,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

JJ Cray Jr., Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #2: Yes

**********

6. Review Comments to the Author

Reviewer #2: There are still some results in the methods section. For example, in lines 168-170 you are reporting the number of incomplete observations. You should perhaps instead report what you did with incomplete observations, for example "observations with missing values in these variables were dropped from the analysis. Where ethnicity was not recorded, it was coded as 'not specified'". And then somewhere in the text of the results indicate that 72 observations were dropped for having missing values (if this is what you did). The results tables should also then have the 551 'not coded' observations for ethnicity.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

PLoS One. doi: 10.1371/journal.pone.0260134.r004

Acceptance letter

JJ Cray Jr

9 Nov 2021

PONE-D-21-13251R1

Dear Dr. Taylor:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. JJ Cray Jr.

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Table. Sensitivity analysis showing the effect of using different minimal adjustment sets on the adjusted odds ratios calculated in Model 1 in the full sample.

(DOCX)

Click here for additional data file.^{(24.1KB, docx)}

Attachment

Submitted filename: Response to reviewers 2021_09_14.docx

Click here for additional data file.^{(942.1KB, docx)}

Data Availability Statement

[pone.0260134.ref001] 1.Cnattingius S. The epidemiology of smoking during pregnancy: Smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine & Tobacco Research. 2004;6(Suppl_2):S125–S40. [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref002] 2.Zdravkovic T, Genbacev O, McMaster MT, Fisher SJ. The adverse effects of maternal smoking on the human placenta: A review. Placenta. 2005;26:S81–S6. doi: 10.1016/j.placenta.2005.02.003 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref003] 3.Kolås T, Nakling J, Salvesen KA. Smoking during pregnancy increases the risk of preterm births among parous women. Acta Obstet Gynecol Scand. 2000;79(8):644–8. [PubMed] [Google Scholar]

[pone.0260134.ref004] 4.Ananth CV, Smulian JC, Demissie K, Vintzileos AM, Knuppel RA. Placental abruption among singleton and twin births in the United States: risk factor profiles. Am J Epidemiol. 2001;153(8):771–8. doi: 10.1093/aje/153.8.771 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref005] 5.Kyrklund-Blomberg NB, Cnattingius S. Preterm birth and maternal smoking: Risks related to gestational age and onset of delivery. American Journal of Obstetrics & Gynecology. 1998;179(4):1051–5. doi: 10.1016/s0002-9378(98)70214-5 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref006] 6.Mitchell EA, Thompson JMD, Robinson E, Wild CJ, Becroft DMO, Clark PM, et al. Smoking, nicotine and tar and risk of small for gestational age babies. Acta Paediatrica. 2002;91(3):323–8. doi: 10.1080/08035250252834003 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref007] 7.Lie RT, Wilcox AJ, Taylor J, Gjessing HK, Saugstad OD, Aabyholm F, et al. Maternal smoking and oral clefts: the role of detoxification pathway genes. Epidemiology (Cambridge, Mass). 2008;19(4):606–15. doi: 10.1097/EDE.0b013e3181690731 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref008] 8.Lange S, Probst C, Rehm J, Popova S. National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis. The Lancet Global Health. 2018;6(7):e769–e76. doi: 10.1016/S2214-109X(18)30223-7 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref009] 9.NHS Digital. Statistics on Women’s Smoking Status at Time of Delivery: England Quarter 3, 2019–20 [accessed 24 March 2020]. https://digital.nhs.uk/data-and-information/publications/statistical/statistics-on-women-s-smoking-status-at-time-of-delivery-england/statistics-on-womens-smoking-status-at-time-of-delivery-england-quarter-3-2019-20.

[pone.0260134.ref010] 10.Rousseaux S, Seyve E, Chuffart F, Bourova-Flin E, Benmerad M, Charles MA, et al. Immediate and durable effects of maternal tobacco consumption alter placental DNA methylation in enhancer and imprinted gene-containing regions. BMC medicine. 2020;18(1):306. doi: 10.1186/s12916-020-01736-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref011] 11.Nakamura A, Francois O, Lepeule J. Epigenetic Alterations of Maternal Tobacco Smoking during Pregnancy: A Narrative Review. Int J Environ Res Public Health. 2021;18(10). doi: 10.3390/ijerph18105083 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref012] 12.Kvalvik LG, Haug K, Klungsøyr K, Morken N-H, DeRoo LA, Skjærven R. Maternal Smoking Status in Successive Pregnancies and Risk of Having a Small for Gestational Age Infant. Paediatric and Perinatal Epidemiology. 2017;31(1):21–8. doi: 10.1111/ppe.12333 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref013] 13.Fleming TP, Watkins AJ, Velazquez MA, Mathers JC, Prentice AM, Stephenson J, et al. Origins of lifetime health around the time of conception: causes and consequences. The Lancet. 2018;391(10132):1842–52. doi: 10.1016/S0140-6736(18)30312-X [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref014] 14.Mathews F, Yudkin P, Smith RF, Neil A. Nutrient intakes during pregnancy: the influence of smoking status and age. Journal of Epidemiology and Community Health. 2000;54(1):17. doi: 10.1136/jech.54.1.17 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref015] 15.Ziauddeen N, Wilding S, Roderick PJ, Macklon NS, Smith D, Chase D, et al. Predicting the risk of childhood overweight and obesity at 4–5 years using population-level pregnancy and early-life healthcare data. BMC medicine. 2020;18(1):105. doi: 10.1186/s12916-020-01568-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref016] 16.Wilding S, Ziauddeen N, Smith D, Roderick P, Chase D, Alwan NA. Are environmental area characteristics at birth associated with overweight and obesity in school-aged children? Findings from the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the south of England. BMC medicine. 2020;18(1):43. doi: 10.1186/s12916-020-01513-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref017] 17.Wilding S, Ziauddeen N, Roderick P, Smith D, Chase D, Macklon N, et al. Are socioeconomic inequalities in the incidence of small-for-gestational-age birth narrowing? Findings from a population-based cohort in the South of England. BMJ Open. 2019;9(7):e026998. doi: 10.1136/bmjopen-2018-026998 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref018] 18.National Institute for Care and Clinical Excellence. Antenatal care for uncomplicated pregnancies 2019 [accessed 30 June 2021]. https://pathways.nice.org.uk/pathways/antenatal-care-for-uncomplicated-pregnancies. [PubMed]

[pone.0260134.ref019] 19.Norris T, Seaton SE, Manktelow BN, Baker PN, Kurinczuk JJ, Field D, et al. Updated birth weight centiles for England and Wales. Arch Dis Child Fetal Neonatal Ed. 2018;103(6):F577–F82. doi: 10.1136/archdischild-2017-313452 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref020] 20.World Health Organisation: Report of a WHO technical consultation on birth spacing: Geneva, Switzerland 13–15 June W2005. World Health Organization.: World Health Organization; 2007.

[pone.0260134.ref021] 21.Textor J, van der Zander B, Gilthorpe MS, Liskiewicz M, Ellison GT. Robust causal inference using directed acyclic graphs: the R package ’dagitty’. Int J Epidemiol. 2016;45(6):1887–94. doi: 10.1093/ije/dyw341 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref022] 22.Textor J. DAGitty—draw and analyze causal diagrams [Website]. http://www.dagitty.net/.

[pone.0260134.ref023] 23.Suttorp MM, Siegerink B, Jager KJ, Zoccali C, Dekker FW. Graphical presentation of confounding in directed acyclic graphs. Nephrology Dialysis Transplantation. 2015;30(9):1418–23. doi: 10.1093/ndt/gfu325 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref024] 24.Textor J. Drawing and Analyzing Causal DAGs with DAGitty [updated 11 January 2020. Manual]. http://dagitty.net/manual-3.x.pdf.

[pone.0260134.ref025] 25.McCowan L, Horgan RP. Risk factors for small for gestational age infants. Best Practice & Research Clinical Obstetrics & Gynaecology. 2009;23(6):779–93. [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref026] 26.R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria: 2020. [Google Scholar]

[pone.0260134.ref027] 27.Dowle M, Srinivasan A. data.table: Extension of ‘data.frame’. R package version 1.13.0. 2020. [Google Scholar]

[pone.0260134.ref028] 28.Wickham H, François R, Henry L, Müller K. dplyr: A Grammar of Data Manipulation. R package version 1.0.2. 2020. [Google Scholar]

[pone.0260134.ref029] 29.Chongsuvivatwong V. epiDisplay: Epidemiological Data Display Package. R package version 3.5.0.1. 2018. [Google Scholar]

[pone.0260134.ref030] 30.Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag; New York. 2016. [Google Scholar]

[pone.0260134.ref031] 31.Wickham H, Miller E. haven: Import and Export ’SPSS’, Stata’ and ’SAS’ Files. 2020.

[pone.0260134.ref032] 32.Revelle W. psych: Procedures for Personality and Psychological Research, Northwestern University, Evanston, Illinois, USA. [Google Scholar]

[pone.0260134.ref033] 33.Wickham H. Reshaping data with the reshape package. Journal of Statistical Software, 21(12). 2007. [Google Scholar]

[pone.0260134.ref034] 34.Wickham H. tidyr: Tidy Messy Data. R package version 1.1.2. 2020. [Google Scholar]

[pone.0260134.ref035] 35.Post A, Gilljam H, Bremberg S, Galanti MR. Maternal smoking during pregnancy: A comparison between concurrent and retrospective self-reports. Paediatric and Perinatal Epidemiology. 2008;22(2):155–61. doi: 10.1111/j.1365-3016.2007.00917.x [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref036] 36.Tran DT, Roberts CL, Jorm LR, Seeho S, Havard A. Change in smoking status during two consecutive pregnancies: a population-based cohort study. BJOG: An International Journal of Obstetrics & Gynaecology. 2014;121(13):1611–20. doi: 10.1111/1471-0528.12769 [DOI] [PubMed] [Google Scholar]

[pone.0260134.ref037] 37.Hauge LJ, Aarø LE, Torgersen L, Vollrath ME. Smoking during consecutive pregnancies among primiparous women in the population-based Norwegian Mother and Child Cohort Study. Nicotine Tob Res. 2013;15(2):428–34. doi: 10.1093/ntr/nts147 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref038] 38.Bauld L, Graham H, Sinclair L, Flemming K, Naughton F, Ford A, et al. Barriers to and facilitators of smoking cessation in pregnancy and following childbirth: literature review and qualitative study. Health Technol Assess. 2017;21(36):1–158. doi: 10.3310/hta21360 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref039] 39.Notley C, Gentry S, Livingstone-Banks J, Bauld L, Perera R, Hartmann-Boyce J. Incentives for smoking cessation. Cochrane Database of Systematic Reviews. 2019(7). [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0260134.ref040] 40.Office for National Statistics. Ethnicity and National Identity in England and Wales: 2011 [accessed 24 March 2020]. https://www.ons.gov.uk/peoplepopulationandcommunity/culturalidentity/ethnicity/articles/ethnicityandnationalidentityinenglandandwales/2012-12-11.

[pone.0260134.ref041] 41.English indices of deprivation 2019 [accessed 1 September 2020]. https://www.gov.uk/government/statistics/english-indices-of-deprivation-2019.

[pone.0260134.ref042] 42.Southampton data observatory [accessed 1 September 2020]. https://data.southampton.gov.uk/economy/deprivation-poverty/.

[pone.0260134.ref043] 43.Dukic VM, Niessner M, Pickett KE, Benowitz NL, Wakschlag LS. Calibrating self-reported measures of maternal smoking in pregnancy via bioassays using a Monte Carlo approach. Int J Environ Res Public Health. 2009;6(6):1744–59. doi: 10.3390/ijerph6061744 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Maternal smoking behaviour across the first two pregnancies and small for gestational age birth: Analysis of the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the South of England

Elizabeth J Taylor

Pia Doh

Nida Ziauddeen

Keith M Godfrey

Ann Berrington

Nisreen A Alwan

Roles

Abstract

Introduction

Methods

Fig 1. Flow diagram showing the composition of the final data used in this analysis.

Assessment of the exposure

Exposure category definitions

Table 1. Summary of derived smoking categories based on self-reported maternal smoking status recorded at the first antenatal appointment for each pregnancy.

Outcome assessment

Assessment of covariates

Statistical analysis

Fig 2. Directed acyclic graph showing the exposure (interpregnancy smoking change) and the outcome (being born small for gestational age (SGA)).

Results

Table 2. Maternal characteristics recorded at the first antenatal appointment at the start of the second pregnancy, together with characteristics of both the first and second infants.

Fig 3. The percentages of small for gestational age births in the first and second pregnancies.

Table 3. Univariate odds of small for gestational age birth (< 10th percentile) in the second pregnancy in the full sample, by maternal characteristics recorded at the start of or during the first and second pregnancies.

Table 4. The odds of small for gestational age birth (< 10th percentile) in the second pregnancy.

Table 5. Full results of Model 1 in Table 4; the adjusted odds of small for gestational age birth (<10th percentile) in the second pregnancy in the full sample.

Discussion

Strengths and limitations

Conclusion

Supporting information

Acknowledgments

Abbreviations

Data Availability

Funding Statement

References

Decision Letter 0

JJ Cray Jr

Roles

Author response to Decision Letter 0

Decision Letter 1

JJ Cray Jr

Roles

Acceptance letter

JJ Cray Jr

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Table 3. Univariate odds of small for gestational age birth (< 10^th percentile) in the second pregnancy in the full sample, by maternal characteristics recorded at the start of or during the first and second pregnancies.

Table 5. Full results of Model 1 in Table 4; the adjusted odds of small for gestational age birth (<10^th percentile) in the second pregnancy in the full sample.