Early Life Exposures

Parental smoking during pregnancy and the risk

of gestational diabetes in the daughter

Wei Bao,1,2 Karin B Michels,3,4,5 Deirdre K Tobias,6,7 Shanshan Li,1

Jorge E Chavarro,3,4,7 Audrey J Gaskins,7 Allan A Vaag,8

Frank B Hu3,4,7 and Cuilin Zhang1*

1Division of Intramural Population Health Research, Eunice Kennedy Shriver National Institute of Child

Health and Human Development, Rockville, MD, USA, 2Department of Epidemiology, University of Iowa

College of Public Health, Iowa City, IA, USA, 3Department of Epidemiology, Harvard T.H. Chan School of

Public Health, Boston, MA, USA, 4Channing Division of Network Medicine, 5Department of Obstetrics,

Gynecology and Reproductive Biology, 6Division of Preventive Medicine, Brigham and Women’s

Hospital and Harvard Medical School, Boston, MA, USA, 7Department of Nutrition, Harvard T.H. Chan

School of Public Health, Boston, MA, USA and 8Department of Endocrinology, Rigshospitalet,

Copenhagen, Denmark

*Corresponding author. Epidemiology Branch, Division of Intramural Population Health Research, Eunice Kennedy

Shriver National Institute of Child Health and Human Development, National Institutes of Health, 6100 Executive Blvd,

Rockville, MD 20852, USA. E-mail: [email protected]

The abstract of this study was included in poster presentation at the American Heart Association’s EPI/Lifestyle2015

Scientific Sessions (March 3–6, 2015, Baltimore, MD).

Accepted 19 November 2015

Abstract

Background: Fetal exposure to parental smoking may have long-term impact on the de-

velopment of disease in adulthood. We examined the association of parental smoking

during pregnancy with risk of gestational diabetes mellitus (GDM) in the daughter.

Methods: We included 15 665 singleton pregnancies from 10 152 women in the Nurses’

Health Study II cohort whose mothers participated in the Nurses’ Mothers’ Cohort Study.

Data on maternal and paternal smoking during pregnancy and associated covariates

were recalled by the mothers. GDM diagnosis was self-reported by the daughters

and was validated by medical record review in a previous study. We used log-binomial

models with generalized estimating equations to estimate relative risks (RRs) and 95%

confidence intervals (CIs).

Results: We observed a positive association between maternal heavy smoking during

pregnancy and risk of GDM in the daughter. The multivariable-adjusted RRs (95% CIs) of

GDM among women whose mothers did not smoke during pregnancy, continued smoking

1–14, 15–24, and�25 cigarettes/day were 1.00 (reference), 1.05 (0.81–1.35), 1.27 (0.95–1.70)and 1.98 (1.18–3.30), respectively (P for trend¼0.01). Further adjustment for the women’sperinatal variables, adult-life characteristics and body mass index during various periods

of life modestly attenuated the association. No association was observed between pater-

nal smoking during the pregnancy period and risk of GDM in the daughter.

Published by Oxford University Press on behalf of International Epidemiological Association 2016.

This work is written by US Government employees and is in the public domain in the US. 160

International Journal of Epidemiology, 2016, 160–169

doi: 10.1093/ije/dyv334

Advance Access Publication Date: 9 January 2016

Original article

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

Conclusions: Maternal heavy smoking (�25 cigarettes/day) during pregnancy was associ-ated with higher risk of gestational diabetes in the daughter. Further studies are warranted

to confirm our findings and to elucidate the underlying mechanisms.

Key words: Gestational diabetes mellitus, maternal smoking during pregnancy

Introduction

The developmental origins of health and disease hypothesis,

or ‘Barker hypothesis’,1 continues to fuel research interest in

examining the health consequences of in utero exposures.

Maternal smoking during pregnancy represents a common

deleterious fetal exposure in many populations.2–4 The

short-term effects of maternal smoking during pregnancy on

multiple adverse pregnancy and perinatal outcomes, includ-

ing fetal growth restriction and low birthweight, have long

been recognized and established.5 Maternal smoking during

pregnancy has also been associated with an increased risk of

obesity during childhood and adulthood in some, although

not all, studies.6–10 There is limited evidence regarding the

long-term impact of fetal exposure to maternal smoking on

the risk of chronic disease in adulthood, which emerges as a

new focus of research interests.11

Gestational diabetes mellitus (GDM) is a common preg-

nancy complication characterized by glucose intolerance,

with onset or first recognition during pregnancy.12 GDM is

not only associated with short-term adverse perinatal out-

comes,13 but also related to long-term metabolic risk in both

mothers and their children.12,14,15 Thus, it is crucial to iden-

tify modifiable risk factors that may contribute to the preven-

tion of GDM in current and subsequent generations.

Animal studies have suggested that fetal exposure to maternal

smoking may lead to impaired glucose metabolism by alter-

ing pancreatic islet development and inducing beta cell apop-

tosis.16,17 In addition, epidemiological studies, although still

limited, suggest that maternal smoking during pregnancy

may increase the risk of diabetes in adulthood.18,19 However

the association, in particular the dose-response relation, be-

tween fetal exposure to maternal smoking and risk of GDM

is not well established. Moreover, no previous study has

examined the association between fetal exposure to paternal

smoking, a major source of maternal passive smoking, and

subsequent risk of GDM. In this study, we aimed to examine

the dose-response relation of maternal and/or paternal smok-

ing during pregnancy with risk of GDM in the daughter.

Methods

Study population

The Nurses’ Health Study II (NHSII) is an ongoing prospect-

ive cohort study of 116 430 female nurses aged 24–44 years

at study inception in 1989. The participants receive a biennial

questionnaire regarding lifestyle behaviours, anthropometric

variables and disease outcomes. In 2001, mothers of the

NHSII participants were invited to complete a questionnaire

regarding their nurse daughter. Details about the Nurses’

Mothers’ Cohort Study have been described elsewhere.20 We

included NHSII participants in the current analyses if they re-

ported at least one singleton pregnancy lasting greater than 6

months between 1989 and 2001 and their mothers partici-

pated in the Nurses’ Mothers’ Cohort Study and reported

data on pregnancy and perinatal variables associated with

the nurse daughter. The NHSII participants were excluded if

they had been adopted, were missing information on mater-

nal smoking or had type 2 diabetes reported in 1989 or be-

fore GDM. Figure 1 depicts the flowchart of study

participants. This study has been approved by the Partners

Human Research Committee (Boston, MA), with partici-

pants’ consent implied by the return of the completed

questionnaires.

Assessment of parental smoking

We used information on parental smoking during pregnancy

from the 2001 Nurses’ Mothers’ Cohort Study question-

naire.21 The mothers reported whether they ever smoked

cigarettes during pregnancy with the nurse daughter, the

number of cigarettes (i.e. 1–14, 15–24, 25–34 or� 35) they

Key Messages

• This study examined the association of parental smoking during pregnancy with risk of gestational diabetes in the

daughter among 15 665 singleton pregnancies from 10 152 women in the Nurses’ Health Study II cohort whose moth-

ers participated in the Nurses’ Mothers’ Cohort Study.

• We demonstrated that maternal heavy smoking (�25 cigarettes/day) during pregnancy was associated with higherrisk of gestational diabetes in the daughter. We did not observe an association between paternal smoking during

pregnancy and risk of gestational diabetes in the daughter.

International Journal of Epidemiology, 2016, Vol. 45, No. 1 161

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

smoked daily during pregnancy, whether they quit smoking

during pregnancy, and if so, in which trimester. In a separ-

ate validation study, the validity of recalled maternal smok-

ing during pregnancy was found to be high in the National

Collaborative Perinatal Project (sensitivity¼0.86, specifi-city¼0.94).22 We categorized maternal smoking as: neversmoked; quit smoking in first trimester of pregnancy; con-

tinued smoking 1–14 cigarettes/day during pregnancy; con-

tinued smoking 15–24 cigarettes/day during pregnancy; and

continued smoking 25 or more cigarettes/day during preg-

nancy. We also asked the mothers whether the nurse’s father

ever smoked during pregnancy and the number of cigarettes

he smoked. We categorized paternal smoking as: never

smoked; smoked 1–14 cigarettes/day during pregnancy;

smoked 15–24 cigarettes/day during pregnancy; and

smoked 25 or more cigarettes/day during pregnancy.

Previous studies based on the same cohort as the current

analysis have found maternal smoking during pregnancy be

associated with an increased risk of overweight and obesity

in the daughter across adolescence and adult life.7

Ascertainment of gestational diabetes

The NHSII participants (i.e. the daughters) reported preva-

lent GDM in 1989 and incident GDM on each biennial

questionnaire through 2001. GDM was not ascertained

after the 2001 questionnaire in the NHSII cohort, because

the majority of NHSII participants had passed reproduct-

ive age by then. In a previous validation study among a

subgroup of the NHSII cohort, 94% of GDM self-reports

were confirmed by medical records.23 In a random sample

of parous women without GDM, 83% reported a glucose

screening test during pregnancy and 100% reported fre-

quent prenatal urine screenings, suggesting a high level of

GDM surveillance in this cohort.23

Covariates assessment

Covariates for maternal, paternal and perinatal characteris-

tics were obtained from the Nurses’ Mothers’ Cohort Study.

The 2001 Nurses’ Mothers’ Cohort Study questionnaire re-

quested data on the daughter’s gestational age at birth, birth-

weight and breastfeeding status, maternal and paternal age

at birth of the daughter, educational level, occupation and

home ownership at the time of the daughter’s birth, maternal

height, maternal pre-pregnancy weight, weight gain during

pregnancy (< 10, 10–14, 15–19, 20–29, 30–40, > 40

pounds; to convert pounds into kilograms, multiply pounds

by the conversion factor 0.453592.), paternal weight, pater-

nal height, maternal consumption of alcoholic beverages dur-

ing pregnancy, and the occurrence of maternal pregnancy

Nurses’ Health Study II cohort (n = 116 430)

Inclusion criteria: pregnancy ≥ 6 months 1989-2001 (NHSII)

(n = 10 862)

Exclusion criteria:• Nurses were adopted (NMS): n = 34• Missing data on parental smoking (NMS): n = 643• T2DM at baseline or before GDM (NHSII): n = 6• Missing main questionnaire (NHSII): n = 33• Missing data on the daughters’ smoking status

(NHSII): n = 76

Analytical population(n = 10 152)

Mothers participated in the Nurses’ Mothers’ Cohort Study

(n = 35 794)

Figure 1. The flowchart of study participants. Exclusion criteria are not mutually exclusive and individual reasons may not total the number of

excluded participants. GDM denotes gestational diabetes mellitus; NHSII, Nurses’ Health Study II; NMS, Nurses’ Mothers’ Cohort Study; T2DM, type

2 diabetes mellitus.

162 International Journal of Epidemiology, 2016, Vol. 45, No. 1

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

complications (gestational diabetes and preeclampsia) during

the pregnancy of the daughter.

Covariates related to the daughters’ characteristics were

obtained from the NHSII questionnaires. The 1989 NHSII

questionnaire assessed the daughters’ age, height, race/ethni-

city and family history of diabetes at baseline. The daugh-

ters’ weight at 18 years old and current weight, parity, and

smoking status (including the number of cigarettes per day)

were self-reported from the 1989 NHSII questionnaire and

were updated with data from each biennial questionnaire

cycle. Self-reported weight was highly correlated with meas-

ured weight (r¼0.97) in a previous validation study.24

Body mass index (BMI) was computed as weight in kilo-

grams divided by height in metres squared. Dietary intake

was collected every 4 years since 1991 using a previously

validated semi-quantitative food frequency question-

naire.25–27 To assess the overall diet quality of the partici-

pants, we derived a diet score, the Alternate Healthy Eating

Index 2010 (AHEI-2010) for each participant, as previously

described.28 The overall AHEI-2010 ranged from 0 to 110

points, with a higher score indicating a better diet quality.

Physical activity was ascertained in 1989, 1991, 1997 and

2001 by frequency of engaging in common recreational

activities, from which metabolic equivalent (MET)-hours

per week were derived. The questionnaire-based estimates

of total physical activity correlated well with detailed activ-

ity diaries in a previous validation study (r¼0.56).29

Cumulative average of physical activity, total energy intake

and AHEI-2010 score were calculated for each individual at

each time period throughout the follow up, to reduce

within-subject variation and represent long-term habitual

diet and physical activity.30

Statistical analysis

We used log-binomial models with generalized estimating

equations to estimate the relative risks (RRs) and 95% confi-

dence intervals (CIs) of GDM for maternal and paternal

smoking, separately and jointly. Generalized estimating equa-

tions allowed us to account for correlations among repeated

observations (pregnancies) contributed by a single participant

(i.e. the nurse daughter). In the multivariable regression mod-

els, we adjusted for: age and race/ethnicity of the daughters

(Model 1); and additionally for maternal and paternal vari-

ables including maternal and paternal age at time of daugh-

ter’s birth, maternal pre-pregnancy BMI, paternal BMI,

maternal weight gain during pregnancy, maternal pregnancy

complications (gestational diabetes, preeclampsia) and mater-

nal alcohol consumption during pregnancy (Model 2); for the

daughters’ perinatal variables including gestational age at

birth, birthweight, and breastfeeding status (Model 3); for

the daughters’ adult life variables including parity, family

history of diabetes, cigarette smoking, total energy intake,

overall diet quality (i.e. Alternate Healthy Eating Index) and

physical activity (Model 4); for the daughters’ BMI at 18

years old (Model 5); and for the daughters’ pre-pregnancy

BMI (Model 6). We mutually adjusted for maternal and pa-

ternal smoking during pregnancy in all these models. Wald

tests were used to assess the differences between maternal

and paternal associations. The daughters’ BMI and other

adult life covariates were updated during the follow-up.

When categorizing each categorical covariate, we created a

category for missing data. We considered Model 3–Model 6

as sensitivity analyses, because the daughter’s perinatal vari-

ables, adult-life variables and adulthood BMI in these models

are potential intermediates or explanatory variables for the

associations of maternal and paternal smoking during preg-

nancy with the risk of GDM. Tests for linear trend were per-

formed across the categories of the number of cigarettes

smoked for mothers who continued to smoke throughout

pregnancy, with non-smoking during pregnancy as the refer-

ence group (for maternal smoking, the test for trend excluded

the category of mothers who quit smoking during preg-

nancy). Statistical analyses were performed using the SAS

statistical software version 9.2 (SAS Institute Inc., Cary, NC)

and the Stata statistical software version 14.0 (StataCorp LP,

College Station, TX).

Results

We included 15 665 singleton pregnancies from 10 152

women in the Nurses’ Health Study II cohort whose mothers

participated in the Nurses’ Mothers’ Cohort Study. Of them,

736 GDM pregnancies were documented. Characteristics of

mothers, fathers and daughters are shown in Table 1 accord-

ing to maternal smoking status during pregnancy. Mothers

who smoked more frequently during pregnancy were more

likely to consume alcoholic beverages during the pregnancy.

The biological fathers of the daughters whose mothers

smoked during pregnancy were also more likely to smoke

during the pregnancy. Women who were exposed to frequent

maternal smoking during pregnancy had a lower birthweight,

were less likely to be breastfed and were heavier and more

likely to smoke in adulthood.

We observed a dose-response relation between in utero

exposure to maternal smoking and risk of GDM (Table 2).

After adjustment for the daughter’s age, race/ethnicity and

maternal and paternal variables, the RRs (95% CIs) of GDM

among women whose mothers did not smoke during preg-

nancy or continued smoking 1–14, 15–24 or � 25 cigarettes/day were 1.00 (reference), 1.05 (0.81–1.35), 1.27

(0.95–1.70) and 1.98 (1.18–3.30), respectively (P for

trend¼0.02). Further adjustment for the daughter’s perinatalvariables and adult life variables, including pre-pregnancy

International Journal of Epidemiology, 2016, Vol. 45, No. 1 163

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

BMI, only slightly changed the association. No association

was observed between paternal smoking during pregnancy

and the risk of GDM (Table 3). Wald tests showed suggestive

statistical evidence that the maternal smoking associations

may differ from the paternal smoking associations with

GDM risk (P¼0.10). We further examined the joint effect ofboth maternal and paternal smoking during pregnancy on

the risk of GDM. Women whose mother or both parents

smoked during pregnancy � 15 cigarettes/day had a higherrisk of GDM (RR 1.43, 95% CI 1.11–1.85), compared with

women whose parents did not smoke during pregnancy or

smoked < 15 cigarettes/day (Supplementary Figure 1, avail-

able as Supplementary data at IJE online).

In an analysis on the joint effect of maternal smoking

during pregnancy and the participants’ smoking during

adulthood, we found that the women who smoked < 15

cigarettes/day and whose mothers smoked � 15 cigarettes/day during pregnancy had an RR (95% CI) of 1.28 (1.00–

1.64) for GDM, compared with neither the mothers nor

the participants smoking � 15 cigarettes/day (Figure 2).We also performed a stratified analysis according to the

daughters’ own smoking status. Among the daughters who

never smoked, the adjusted RRs (95% CIs) of GDM were

1.30 (0.79–2.13), 0.94 (0.67–1.32), 1.13 (0.77–1.65) and

2.15 (1.19–3.88) for the daughters whose mothers smoked

but quit smoking in the first trimester, continued smoking

Table 1. Age-standardized maternal, paternal, and the daughter’s characteristics by maternal smoking status during pregnancya

Maternal cigarette smoking during pregnancy

Non-smoker Quit smoking in

the first trimester

Continued smoking

1–14 cigarettes/day

Continued

smoking 15–24

cigarettes/day

Continued smoking

�25 cigarettes/day

Number of participants 7478 373 1369 793 139

Maternal characteristics

Age at daughter’s birth (years) 26.68 (4.83) 25.33 (4.34) 26.39 (4.75) 26.23 (4.66) 26.90 (4.65)

Prepregnancy BMI (kg/m2) 21.41 (2.60) 21.00 (2.26) 20.88 (2.46) 21.09 (2.70) 21.24 (2.61)

Attended college (%) 40.21 47.11 43.04 40.89 46.57

Ever consumed alcoholic beverages

during pregnancy (%)

25.03 44.82 60.43 63.59 63.68

Pregnancy complicationsb (%) 4.03 4.26 3.90 3.84 3.44

Paternal characteristics

Age at daughter’s birth (years) 28.70 (4.78) 27.85 (4.68) 28.59 (4.67) 28.57 (4.64) 29.63 (4.52)

BMI at daughter’s birth (kg/m2) 23.80 (2.81) 23.86 (3.08) 23.78 (2.80) 23.72 (2.83) 23.40 (2.75)

Attended college (%) 45.90 54.68 51.51 48.45 57.90

Ever smoked during pregnancy (%) 44.61 71.72 73.86 78.26 74.26

Characteristics of the daughter in early life

Gestational age at birth (weeks) 39.42 (2.26) 39.55 (2.32) 39.32 (2.40) 39.12 (2.54) 39.25 (2.54)

Birthweight (g) 3358.28 (493.41) 3291.89 (491.44) 3176.63 (498.50) 3070.21 (514.39) 3056.42 (523.58)

Caucasian (%) 95.05 95.42 96.03 95.84 95.86

Breastfed during infancy (%) 43.46 47.81 32.07 34.65 23.07

Characteristics of the daughter

during adulthoodc

Age in 1989 (years) 30.62 (3.42) 30.57 (3.27) 30.85 (3.42) 30.33 (3.31) 30.40 (3.29)

BMI (kg/m2) 22.81 (3.93) 23.01 (4.01) 22.87 (3.64) 23.32 (4.13) 23.64 (4.66)

Nulliparous (%) 7.68 7.50 6.02 8.21 5.81

Family history of diabetes (%) 10.08 8.51 8.60 7.55 7.88

Current smoking (%) 6.54 11.06 10.86 11.36 10.95

Alcohol intake (g/day) 2.55 (4.67) 3.15 (4.79) 3.42 (5.74) 3.30 (5.66) 3.33 (6.12)

Physical activity (MET-h/week) 25.88 (36.49) 26.26 (38.56) 27.59 (37.17) 28.19 (44.22) 28.31 (34.78)

Total energy intake (kcal/day) 1882.42 (545.50) 1837.95 (522.48) 1871.40 (549.03) 1850.16 (540.46) 1873.85 (559.33)

AHEI-2010d 47.42 (10.67) 48.72 (10.70) 48.38 (11.09) 48.08 (10.56) 47.85 (10.75)

AHEI-2010 indicates Alternate Healthy Eating Index 2010; BMI, body mass index; MET, metabolic equivalent.aValues are means (standard deviations) for continuous variables and percentages for categorical variables and are standardized to age distribution of the

NHSII participants (i.e. the daughters).bMaternal pregnancy complications included gestational diabetes and preeclampsia.cAdulthood characteristics are provided for 1989, except diet information (i.e. total energy intake, alcohol intake and the derived alternate healthy eating

index) which was first collected in the Nurses’ Health Study II cohort in 1991.dAHEI-2010 was derived for each participant, as previously describe,28 to assess the overall diet quality of the participants.

164 International Journal of Epidemiology, 2016, Vol. 45, No. 1

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

Tab

le2.M

ate

rna

lcig

are

tte

sm

okin

gd

uri

ng

pre

gn

an

cy

an

dth

eri

sk

of

ge

sta

tio

na

ld

iab

ete

sin

the

da

ug

hte

r

Mate

rnalcig

are

tte

smokin

gduri

ng

pre

gnancy

Pfo

rtr

end

d

Non-s

mokerQ

uit

smokin

gin

the

firs

ttr

imest

erC

onti

nued

smokin

g

1–14

cig

are

ttes/

day

Conti

nued

smokin

g

15–24

cig

are

ttes/

day

Conti

nued

smokin

g

‡25

cig

are

ttes/

day

GD

Mca

ses/

pre

gnanci

es522/1

1525

33/5

59

95/2

122

68/1

238

18/2

21

Model

1:A

dju

sted

for

the

daughte

r’s

age

and

race

/eth

nic

ity

1.0

01.2

6(0

.84–1.8

7)1.0

0(0

.79–1.2

8)

1.2

2(0

.92–1.6

2)

1.8

6(1

.12–3.0

9)

0.0

2

Model

2:M

odel

1þ

addit

ionally

adju

sted

for

mate

rnaland

pate

rnalvari

able

sa1.0

01.2

7(0

.85–1.8

9)1.0

5(0

.81–1.3

5)

1.2

7(0

.95–1.7

0)

1.9

8(1

.18–3.3

0)

0.0

1

Model

3:M

odel

2þ

addit

ionally

adju

sted

for

the

daughte

r’s

per

inata

lvari

able

sb1.0

01.2

5(0

.84–1.8

6)1.0

1(0

.78–1.3

0)

1.2

1(0

.90–1.6

3)

1.8

9(1

.13–3.1

5)

0.0

4

Model

4:M

odel

3þ

addit

ionally

adju

sted

for

the

daughte

r’s

adult

-lif

ech

ara

cter

isti

csc

1.0

01.2

8(0

.87–1.9

0)1.0

1(0

.78–1.3

0)

1.2

3(0

.91–1.6

5)

1.9

7(1

.20–3.2

4)

0.0

2

Model

5:M

odel

4þ

addit

ionally

adju

sted

for

the

daughte

r’s

BM

Iat

age

18

yea

rs1.0

01.2

7(0

.86–1.8

7)1.0

0(0

.77–1.2

8)

1.2

2(0

.91–1.6

4)

1.8

8(1

.15–3.0

7)

0.0

3

Model

6:M

odel

5þ

addit

ionally

adju

sted

for

the

daughte

r’s

most

rece

nt

pre

-pre

gnancy

BM

I1.0

01.2

4(0

.85–1.8

1)1.0

0(0

.78–1.2

9)

1.1

7(0

.88–1.5

6)

1.8

5(1

.12–3.0

4)

0.0

5

Pate

rnalsm

okin

gduri

ng

pre

gnancy

was

adju

sted

inall

the

model

s.aT

he

mate

rnaland

pate

rnalvari

able

sin

cluded

mate

rnaland

pate

rnalage

at

tim

eof

the

daughte

r’s

bir

th,m

ate

rnalpre

-pre

gnancy

body

mass

index

,pate

rnalbody

mass

index

,m

ate

rnalw

eight

gain

duri

ng

pre

gnancy

,m

ate

r-

nalpre

gnancy

com

plica

tions

(ges

tati

onaldia

bet

es,pre

ecla

mpsi

a),

and

mate

rnalco

nsu

mpti

on

of

alc

oholic

bev

erages

duri

ng

pre

gnancy

.T

he

info

rmati

on

was

report

edin

the

Nurs

es’M

oth

ers

Stu

dy

ques

tionnair

e.

bT

he

per

inata

lvari

able

sin

cluded

the

daughte

r’s

ges

tati

onalage

at

bir

th,bir

thw

eight

and

bre

ast

feed

ing

statu

s.T

he

info

rmati

on

was

report

edin

the

Nurs

es’M

oth

ers

Stu

dy

ques

tionnair

e.

cT

he

daughte

rs’adult

-lif

evari

able

sin

cluded

pari

ty,fa

mily

his

tory

of

dia

bet

es,ci

gare

tte

smokin

g,physi

calact

ivit

y,to

talen

ergy

inta

ke,

and

over

all

die

tquality

(i.e

.A

lter

nate

Hea

lthy

Eati

ng

Index

).T

he

info

rmati

on

was

re-

port

edin

the

Nurs

es’H

ealt

hStu

dy

IIques

tionnair

e.dP

for

tren

dacr

oss

non-s

moker

s,sm

okin

g1–14

cigare

ttes

/day,sm

okin

g15–24

cigare

ttes

/day

and

smokin

g�

15

cigare

ttes

/day.

Tab

le3.P

ate

rna

lcig

are

tte

sm

okin

gd

uri

ng

pre

gn

an

cy

an

dth

eri

sk

of

ge

sta

tio

na

ld

iab

ete

sin

the

da

ug

hte

r

Pate

rnalci

gare

tte

smokin

gduri

ng

pre

gnancy

Pfo

rtr

end

Non-s

moker

1–14

cigare

ttes

/day

15–24

cigare

ttes

/day

�25

cigare

ttes

/day

GD

Mca

ses/

tota

lpart

icip

ants

350/7

527

128/3

071

166/3

433

92/1

634

Model

1:A

dju

sted

for

the

daughte

r’s

age

and

race

/eth

nic

ity

1.0

00.9

1(0

.73–1.1

3)

1.0

2(0

.83–1.2

6)

1.1

9(0

.92–1.5

4)

0.2

6

Model

2:M

odel

1þ

addit

ionally

adju

sted

for

mate

rnaland

pate

rnalvari

able

sa1.0

00.9

2(0

.74–1.1

5)

1.0

2(0

.82–1.2

5)

1.1

6(0

.90–1.5

1)

0.3

2

Model

3:M

odel

2þ

addit

ionally

adju

sted

for

the

daughte

r’s

per

inata

lvari

able

sb1.0

00.9

1(0

.73–1.1

3)

1.0

1(0

.82–1.2

4)

1.1

4(0

.88–1.4

8)

0.3

8

Model

4:M

odel

3þ

addit

ionally

adju

sted

for

the

daughte

r’s

adult

-lif

ech

ara

cter

isti

csc

1.0

00.9

1(0

.73–1.1

4)

1.0

0(0

.81–1.2

2)

1.0

9(0

.84–1.4

1)

0.5

9

Model

5:M

odel

4þ

Addit

ionally

adju

sted

for

the

daughte

r’s

BM

Iat

age

18

yea

rs1.0

00.9

1(0

.73–1.1

3)

0.9

8(0

.80–1.2

1)

1.0

7(0

.82–1.3

9)

0.6

9

Model

6:M

odel

5þ

addit

ionally

adju

sted

for

the

daughte

r’s

most

rece

nt

pre

-pre

gnancy

BM

I1.0

00.9

2(0

.74–1.1

5)

0.9

9(0

.80–1.2

1)

1.0

3(0

.80–1.3

3)

0.8

6

BM

Iden

ote

sbody

mass

index

.

Mate

rnalsm

okin

gduri

ng

pre

gnancy

was

adju

sted

inall

the

model

s.aT

he

mate

rnal

and

pate

rnal

per

inata

lvari

able

sin

cluded

mate

rnal

and

pate

rnal

age

at

tim

eof

the

daughte

r’s

bir

th,

mate

rnal

pre

-pre

gnancy

body

mass

index

,pate

rnal

body

mass

index

,m

ate

rnal

wei

ght

gain

duri

ng

pre

g-

nancy

,m

ate

rnalpre

gnancy

com

plica

tions

(ges

tati

onaldia

bet

es,pre

ecla

mpsi

a),

and

mate

rnalco

nsu

mpti

on

of

alc

oholic

bev

erages

duri

ng

pre

gnancy

.T

he

info

rmati

on

was

report

edin

the

Nurs

es’M

oth

ers

Stu

dy

ques

tionnair

e.bT

he

per

inata

lvari

able

sin

cluded

the

daughte

r’s

ges

tati

onalage

at

bir

th,bir

thw

eight

and

bre

ast

feed

ing

statu

s.T

he

info

rmati

on

was

report

edin

the

Nurs

es’M

oth

ers

Stu

dy

ques

tionnair

e.cT

he

daughte

rs’adult

-lif

evari

able

sin

cluded

pari

ty,fa

mily

his

tory

of

dia

bet

es,ci

gare

tte

smokin

g,physi

calact

ivit

y,to

talen

ergy

inta

ke

and

over

all

die

tquality

(i.e

.A

lter

nate

Hea

lthy

Eati

ng

Index

).T

he

info

rmati

on

was

re-

port

edin

the

Nurs

es’H

ealt

hStu

dy

IIques

tionnair

e.

International Journal of Epidemiology, 2016, Vol. 45, No. 1 165

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

1–14, 15–24 or � 25 cigarettes/day, respectively, com-pared with those whose mothers did not smoke during

pregnancy.

Discussion

We observed a positive association between maternal

smoking during pregnancy and risk of GDM in the daugh-

ter. Specifically, maternal smoking of � 25 cigarettes/dayduring pregnancy was associated with 98% higher risk of

GDM in the daughter. The association was independent of

other major risk factors during and after pregnancy and it

was only slightly changed after adjustment for the daugh-

ter’s birthweight and adult life variables including pre-

pregnancy BMI. We did not find an association between

paternal smoking and GDM in the daughter.

With detailed information on smoking exposure as well as

covariates during and after the pregnancy, our study expands

previous findings on maternal smoking during pregnancy and

GDM risk. The MoBa Cohort in Norway recently reported

that in utero exposure to maternal tobacco smoke was associ-

ated with increased risk of GDM.8 However, the dose-re-

sponse relation between maternal smoking during pregnancy

and GDM risk in the daughter was not assessed in that study

(i.e. maternal smoking was assessed by yes versus no, without

information on the dose). Furthermore, other maternal vari-

ables were not available in that study, which may have lim-

ited its ability to evaluate the impact of potential confounders

on the observed association. A subsequent study using data

from the Swedish Medical Birth Register yielded similar find-

ings in a younger population (age range 13–28 years, with

70% � 24 years).9 However, the generalizability of resultsamong this younger population may be limited, given that

the incidence of GDM is higher in women aged 30 years or

older.31 In addition, the registry had relatively limited infor-

mation on maternal covariates. In the present study, detailed

information on maternal and paternal characteristics was col-

lected in the Nurses’ Mothers’ Cohort Study. Thus, the com-

bination of parental data with the daughter’s data collected

in the NHSII provides a unique opportunity to examine the

long-term intergenerational impact of parental smoking on

adulthood diseases in the daughter. Our results were also in

line with previous studies regarding the association between

maternal smoking during pregnancy and risk of type 2 dia-

betes in adulthood.18,19

The observed association between maternal heavy

smoking during pregnancy and higher GDM risk is bio-

logically plausible. Maternal smoking, as a deleterious in

utero environmental insult, may lead to structural, physio-

logical and metabolic changes to the fetus and result in

0.5

11

.52

Rel

ativ

e ri

sk o

f ges

tatio

nal d

iabe

tes

-/- +/- -/+ or +/+Maternal/the daughter’s smoking status

Figure 2. Joint effect of maternal smoking during pregnancy and the daughter’s smoking during adulthood on the risk of GDM in the daughter. The

symbol ‘þ’ indicates smoking � 15 cigarettes/day, ‘-’ indicates no smoking or smoking < 15 cigarettes/day. The symbol before and after the ‘/’ de-notes smoking status of the mothers and the daughters, respectively. Covariates included the daughters’ age, maternal and paternal age at time of

the daughter’s birth, maternal pre-pregnancy body mass index, paternal body mass index, maternal weight gain during pregnancy, maternal preg-

nancy complications (gestational diabetes, preeclampsia), maternal consumption of alcoholic beverages during pregnancy, paternal smoking during

pregnancy, daughter’s race/ethnicity, gestational age at birth, breastfeeding status, birth eight, parity, family history of diabetes, physical activity,

total energy intake, overall diet quality (i.e. Alternate Healthy Eating Index), BMI at 18 years old and updated adulthood BMI.

166 International Journal of Epidemiology, 2016, Vol. 45, No. 1

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

impaired glucose metabolism and metabolic diseases in

adulthood.18,32 Animal studies suggest that fetal exposure

to smoking might lead to altered pancreatic islet and adi-

pose tissue development, beta cell apoptosis at birth and

postnatal endocrine and metabolic changes.16,17 The

observed divergent associations of maternal and paternal

smoking during pregnancy with GDM risk in the daughter

may indicate a specific intrauterine effect of maternal

smoking, rather than shared familial confounding fac-

tors,33 on GDM risk. This is in line with findings from a re-

cent meta-analysis showing greater effect estimates of

maternal smoking in pregnancy than paternal smoking in

association with childhood obesity in the offspring.34

There has been little evidence regarding in which tri-

mester of pregnancy fetal exposure to smoking most influ-

ences pancreatic islet and other metabolic effects that may

increase the risk of GDM or type 2 diabetes. In the current

study, we observed a dose-response relation between con-

tinued maternal smoking during pregnancy and GDM risk

in the daughter. We also observed suggestive evidence that

the daughters of women who smoked only in the first tri-

mester had a higher risk of GDM. Coincidentally, a previ-

ous study found that fetal exposure to maternal smoking

during the first trimester only was associated with type 2

diabetes in adulthood.19 These results indicated that the

first trimester of pregnancy might be a sensitive window

for the development of diabetes induced by maternal

smoking, which warrants confirmation in future studies.

Our study had several strengths, including the large

sample size, the availability of detailed information on

both maternal and paternal smoking during pregnancy

that allows analyses on dose-response relation, and the

comprehensive information on potential confounders col-

lected in the Nurses’ Mothers’ Cohort Study. Moreover,

the study relied primarily on maternal reports of parental

smoking during pregnancy, which is likely to be less mis-

classified than reports by the daughters.

We acknowledge that there were several limitations. First,

the ascertainment of GDM was based on self-reports in the

NHSII cohort. A previous validation study in a subset of this

cohort found a high validity (94%) of GDM self-reports

compared with medical record reviews,23 reducing the con-

cern of outcome misclassification. Second, data on parental

smoking and associated covariates during pregnancy were

collected retrospectively in the Nurses’ Mothers’ Cohort

Study. Although the validity of recalled maternal smoking

during pregnancy was found to be high (sensitivity¼0.86,specificity¼0.94) in a similar study,22 it is possible thatmothers who smoked during pregnancy were more likely to

be misclassified as not smoking than non-smoking mothers

were to be misclassified as smokers. Such differential mis-

classification would lead to an underestimation of the true

association of maternal smoking on daughter’s GDM risk in

this study. In addition, the mothers, in particular those who

had other children besides the nurse daughter, might misre-

member the circumstances (including parental smoking sta-

tus) surrounding the pregnancy of the nurse daughter a long

time after their pregnancies. It would be even more challeng-

ing for the mothers to recall paternal smoking status during

the specific pregnancy with the nurse daughter. However,

since the data about parental smoking during pregnancy

were collected without reference to the daughter’s GDM sta-

tus in the present study, any misclassification due to misre-

membering the parental smoking status during pregnancy

would be non-differential with respect to GDM outcome in

the daughter, which may also lead to an underestimation of

the true association. Therefore, the true detrimental effects

of maternal smoking on daughter’s GDM risk would be even

stronger if misclassification bias could be minimized. Third,

we did not have information about maternal smoking before

pregnancy. Mothers who quit smoking before pregnancy

were combined with never smokers in the reference group,

which may underestimate the true effect of maternal smok-

ing during pregnancy on the daughter’s risk of GDM.

Fourth, the Nurses’ Mothers’ Cohort was limited to mothers

who were alive and able to complete the questionnaire in

2001. At that time, half of the mother participants were over

70 years old. Since smoking is associated with higher risk of

mortality and various morbidities, mothers who had smok-

ing habits were less likely to survive or be able to participate

in this cohort. As a result, the survival effect may underesti-

mate the true association in this study. Fifth, our study popu-

lation consisted mostly of Caucasian American women.

Future research among other race/ethnic groups is needed.

However, the relative homogeneity of the study population

reduces potential confounding due to unmeasured socio-eco-

nomic variability. Finally, although we have considered

many potential confounders in this analysis, we cannot com-

pletely exclude the possibilities of residual confounding from

unmeasured factors. For instance, genetic factors may also

confound the association in addition to other factors. It has

been demonstrated that a genetic variant related to smoking

behaviour is associated with adiposity, which will in turn be

associated with GDM, even among people who have never

smoked.35

In conclusion, maternal heavy smoking (� 25 cigarettes/day) during pregnancy was associated with higher risk of

GDM in the daughter. This study expanded our knowledge

on the adverse health effects of maternal smoking during

pregnancy, not only leading to short-term adverse pregnancy

and perinatal outcomes but also increasing the long-term

intergenerational risk of GDM in the daughter. It adds evi-

dence to support the recommendation that maternal smoking

during pregnancy should be strongly discouraged. Further

International Journal of Epidemiology, 2016, Vol. 45, No. 1 167

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

studies are warranted to confirm our findings and to eluci-

date the underlying mechanisms.

Supplementary Data

Supplementary data are available at IJE online.

Funding

This study was supported by the Intramural Research Program of

the Eunice Kennedy Shriver National Institute of Child Health and

Human Development, National Institutes of Health (contract No.

HHSN275201000020C). The Nurses’ Health Study II was funded

by research grants DK58845, CA50385, P30 DK46200 and UM1

CA176726 from the National Institutes of Health. The Nurses’

Mothers’ Cohort Study was funded by the Intramural Research

Program of the National Cancer Institute, Research Contract N02-

RC-17027 from the National Cancer Institute, and by P.O. 263 MQ

411027 from the National Cancer Institute. D.T. was supported by

a mentored fellowship from the American Diabetes Association

(No. 7-12-MN-34) and a K01 grant from National Institute of

Diabetes and Digestive and Kidney Diseases (DK58845). A.G. was

supported by a training grant from National Institute of Diabetes

and Digestive and Kidney Diseases (T32-DK007703-16).

Contributors

W.B. and C.Z. conceived the idea and designed the study.

W.B. wrote the manuscript. K.B.M., D.K.T., S.L., J.E.C,

A.J.G., A.A.V., F.B.H. and C.Z. interpreted the results and

reviewed and edited the manuscript. S.L. conducted tech-

nique review for this manuscript. W.B. and C.Z. had pri-

mary responsibility for final content. All authors provided

intellectual input into the paper, and all authors read and

approved the final manuscript.

Conflict of interest: All the authors declare no conflicts of

interest.

References

1. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of

in utero and early-life conditions on adult health and disease. N

Engl J Med 2008;359:61–73.

2. Cnattingius S. The epidemiology of smoking during pregnancy:

smoking prevalence, maternal characteristics, and pregnancy

outcomes. Nicotine Tob Res 2004;6(Suppl 2):S125–40.

3. Tong VT, Dietz PM, Morrow B et al. Trends in smoking before,

during, and after pregnancy—Pregnancy Risk Assessment

Monitoring System, United States, 40 sites, 2000–2010. MMWR

Surveill Summ 2013;62:1–19.

4. Gilbert NL, Bartholomew S, Raynault MF, Kramer MS.

Temporal Trends in Social Disparities in Maternal Smoking and

Breastfeeding in Canada, 1992–2008. Matern Child Health J

2014;18:1905–11.

5. Rogers JM. Tobacco and pregnancy. Reprod Toxicol 2009;

28:152–60.

6. Durmus B, Kruithof CJ, Gillman MH et al. Parental smoking

during pregnancy, early growth, and risk of obesity in preschool

children: the Generation R Study. Am J Clin Nutr 2011; 94:

164–71.

7. Harris HR, Willett WC, Michels KB. Parental smoking during

pregnancy and risk of overweight and obesity in the daughter.

Int J Obes (Lond) 2013;37:1356–63.

8. Cupul-Uicab LA, Skjaerven R, Haug K, Melve KK, Engel SM,

Longnecker MP. In utero exposure to maternal tobacco smoke

and subsequent obesity, hypertension, and gestational diabetes

among women in the MoBa cohort. Environ Health Perspect

2012;120:355–60.

9. Mattsson K, Kallen K, Longnecker MP, Rignell-Hydbom A,

Rylander L. Maternal smoking during pregnancy and daughters’

risk of gestational diabetes and obesity. Diabetologia 2013;

56:1689–95.

10. Oken E, Levitan EB, Gillman MW. Maternal smoking during

pregnancy and child overweight: systematic review and meta-

analysis. Int J Obes (Lond) 2008;32:201–10.

11. Behl M, Rao D, Aagaard K et al. Evaluation of the association

between maternal smoking, childhood obesity, and metabolic

disorders: a national toxicology program workshop review.

Environ Health Perspect 2013;121:170–80.

12. American Diabetes Association. Gestational diabetes mellitus.

Diabetes Care 2004;27(Suppl 1):S88–90.

13. Metzger BE, Lowe LP, Dyer AR et al. Hyperglycemia and

adverse pregnancy outcomes. N Engl J Med 2008;358:1991–

2002.

14. Reece EA, Leguizamon G, Wiznitzer A. Gestational diabetes: the

need for a common ground. Lancet 2009;373:1789–97.

15. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes

mellitus after gestational diabetes: a systematic review and meta-

analysis. Lancet 2009;373:1773–79.

16. Holloway AC, Lim GE, Petrik JJ, Foster WG, Morrison KM,

Gerstein HC. Fetal and neonatal exposure to nicotine in Wistar

rats results in increased beta cell apoptosis at birth and postnatal

endocrine and metabolic changes associated with type 2 dia-

betes. Diabetologia 2005;48:2661–66.

17. Somm E, Schwitzgebel VM, Vauthay DM et al. Prenatal nicotine

exposure alters early pancreatic islet and adipose tissue develop-

ment with consequences on the control of body weight and

glucose metabolism later in life. Endocrinology 2008;149:

6289–99.

18. Montgomery SM, Ekbom A. Smoking during pregnancy and dia-

betes mellitus in a British longitudinal birth cohort. BMJ 2002;

324:26–27.

19. Jaddoe VW, de Jonge LL, van Dam RM et al. Fetal exposure to

parental smoking and the risk of type 2 diabetes in adult women.

Diabetes Care 2014;37:2966–73.

20. Michels KB, Willett WC, Graubard BI et al. A longitudinal study

of infant feeding and obesity throughout life course. Int J Obes

(Lond) 2007;31:1078–85.

21. Simard JF, Rosner BA, Michels KB. Exposure to cigarette smoke

in utero: comparison of reports from mother and daughter.

Epidemiology 2008;19:628–33.

22. Tomeo CA, Rich-Edwards JW, Michels KB et al. Reproducibility

and validity of maternal recall of pregnancy-related events.

Epidemiology 1999;10:774–77.

168 International Journal of Epidemiology, 2016, Vol. 45, No. 1

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020

23. Solomon CG, Willett WC, Rich-Edwards J et al. Variability in

diagnostic evaluation and criteria for gestational diabetes.

Diabetes Care 1996;19:12–16.

24. Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB,

Willett WC. Validity of self-reported waist and hip circumfer-

ences in men and women. Epidemiology 1990;1:466–73.

25. Willett WC, Sampson L, Browne ML et al. The use of a self-

administered questionnaire to assess diet four years in the past.

Am J Epidemiol 1988;127:188–99.

26. Willett WC, Sampson L, Stampfer MJ et al. Reproducibility and

validity of a semiquantitative food frequency questionnaire. Am

J Epidemiol 1985;122:51–65.

27. Salvini S, Hunter DJ, Sampson L et al. Food-based validation of

a dietary questionnaire: the effects of week-to-week variation in

food consumption. Int J Epidemiol 1989;18:858–67.

28. Chiuve SE, Fung TT, Rimm EB et al. Alternative dietary indices

both strongly predict risk of chronic disease. J Nutr 2012;

142:1009–18.

29. Wolf AM, Hunter DJ, Colditz GA et al. Reproducibility and val-

idity of a self-administered physical activity questionnaire. Int J

Epidemiol 1994;23:991–99.

30. Hu FB, Stampfer MJ, Rimm E et al. Dietary fat and coronary

heart disease: a comparison of approaches for adjusting for total

energy intake and modeling repeated dietary measurements. Am

J Epidemiol 1999;149:531–40.

31. Makgoba M, Savvidou MD, Steer PJ. An analysis of the inter-

relationship between maternal age, body mass index and racial

origin in the development of gestational diabetes mellitus. BJOG

2012;119:276–82.

32. Bakker H, Jaddoe VW. Cardiovascular and metabolic in-

fluences of fetal smoke exposure. Eur J Epidemiol 2011;26:

763–70.

33. Davey Smith G. Negative control exposures in epidemiologic

studies. Epidemiology 2012;23:350–51; author reply 51–52.

34. Riedel C, Schonberger K, Yang S et al. Parental smoking and

childhood obesity: higher effect estimates for maternal

smoking in pregnancy compared with paternal smoking—–a

meta-analysis. Int J Epidemiol 2014;43:1593–606.

35. Taylor AE, Morris RW, Fluharty ME et al. Stratification by

smoking status reveals an association of CHRNA5-A3-B4 geno-

type with body mass index in never smokers. PLoS Genet

2014;10:e1004799.

International Journal of Epidemiology, 2016, Vol. 45, No. 1 169

Dow

nloaded from https://academ

ic.oup.com/ije/article-abstract/45/1/160/2363826 by U

niversity of Texas at Tyler user on 10 February 2020