Eating Behaviors 15 (2014) 403–409

Contents lists available at ScienceDirect

Eating Behaviors

The impact of maternal overweight and emotion regulation on earlyeating behaviors

Gaia de Campora a,⁎, Luciano Giromini b, Giovanni Larciprete c, Valentina Li Volsi a, Giulio Cesare Zavattini a

a Sapienza University of Rome, Department of Dynamic and Clinical Psychology, Via degli Apuli 1, 00185, Italyb Alliant International University, California School of Professional Psychology, 10455 Pomerado Road, San Diego, CA 92131, United Statesc Fatebenefratelli Hospital, Department of Obstetrics and Gynecology, Piazza Fatebenefratelli 2, Rome, 00186, Italy

⁎ Corresponding author. Tel.: +39 340 2996770.E-mail addresses: gaiadecampora@gmail.com (G. de C

(L. Giromini), giovanni.larciprete@fbf-isola.it (G. Larciprete)giuliocesare.zavattini@uniroma1.it (G.C. Zavattini).

1 This research is part of a larger and ongoing longitudi

http://dx.doi.org/10.1016/j.eatbeh.2014.04.0131471-0153/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 7 October 2013Received in revised form 2 March 2014Accepted 29 April 2014Available online 10 May 2014

Keywords:Emotion regulationEarly eating behaviorsPregnancyBMIOverweightRisk factors

1Empirical data indicate that the risk for childhood obesity and overweight increases when one or both parentsare overweight or obese. Such an association, however, cannot be entirely explained only by biological factors.Based on available literature, we hypothesized that maternal emotion regulation might play a role in explainingthe intergenerational transfer of overweight and obesity. We conducted a quasi-experimental, longitudinalstudy: (step I) during the third trimester of pregnancy of 65 Italian women (33 overweight and 32 non-overweight), theDifficulties in Emotion Regulation Scalewere administered to assess the quality of their emotionregulation strategies; and (step II) seven months after the delivery, the feeding interactions between the partic-ipants and their babies were evaluated in a 20-minute video-recording, by using the Italian version of the Obser-vational Scale for Mother–Infant Interaction during Feeding. When compared to the non-overweight group, theoverweight group had more difficulties in emotion regulation, was more psychologically distressed, and hadpoorer feeding interactions with their babies. Perhaps more importantly, the extent to which the participantswere suffering difficulties in emotion regulation during pregnancy predicted, significantly, and beyond the effectsof pre-pregnancy maternal weight, the quality of the mother–child feeding interactions 7 months after thedelivery.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Overweight and obesity consist of an imbalance between calorieintake and expenditure. Overweight individuals have a Body MassIndex (BMI; weight in kilograms divided by the square of the height inmeters) between 25 and 30 while obese individuals have a BMI greaterthan 30. These rapidly increasing conditions are primarily diet-induced,resulting from sustained excess of energy dense, high fat, and refinedcarbohydrate content foods, aswell as insufficient consumption of fruitsand vegetables. The increasingly sedentary lifestyles and changing envi-ronmentswhich restrict opportunities for physical activity, also contrib-ute to their development.

Despite the high prevalence of these phenomena, to date the rela-tionship between weight and psychological health remains controver-sial and poorly understood. A number of risk factors for overweightand obesity have been linked to demographic aspects, dietary habits,

ampora), lgiromini@alliant.edu, vale.liv@hotmail.it (V. Li Volsi),

nal investigation.

social/environmental and cognitive factors (Van der Merwe, 2007).However, the specific psychological mechanisms through which theserisk factors affect the attitude toward food and lifestyle, and conse-quently behavior and weight, have not been completely clarified.

Most of the research efforts are currently directed toward addressingthe complex etiology underlying these conditions, by integrating genet-ic, physiological and psychological components. A growing body of re-search, in particular, shows that a central role in the development ofobesity and overweight might be played by the parent–child relation-ship. Indeed, the risk among children to be overweight when one orboth parents are overweight or obese dramatically increases as com-pared to peers from non-obesogenic environments, and such an associ-ation cannot be entirely explained by biological factors alone. Manyresearchers, indeed, have demonstrated that factors such as breast feed-ing duration (Agras & Mascola, 2005), use and length of bottle feeding,smoking during pregnancy (Owen, Martin, Whincup, Smith, & Cook,2005), parental style, and their modeling of eating behaviors constitutemajor risk factors in promoting overweight during childhood and inlater ages (Frankel et al., 2012). Furthermore, it has been reported thatthe provision, or not, of the emotional context of the feeding interactionwith the baby (as being permissive or demanding, available or poorlytuned), strongly affects the eating habit of the child (Farrow & Blissett,2008; Ventura & Birch, 2008).

404 G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

1.1. Emotion regulation and early feeding interactions

The caregiving system plays an important role in promoting thephysical, emotional and social wellbeing of the baby. The ability to reg-ulate emotions and internal states during the early stages of life is alsolinked to the caregiver's capacity to handle the emotional states of thechild. A central role in this process, in particular, is played by what hasbeen referred to as maternal sensitivity, i.e., the ability to appropriately,timely, and consistently respond to the baby's signals and needs. AsSchore (2000) pointed out, indeed, during thefirst year of life themoth-er–infant dyad goes on as a mutually regulating biological unit, and thetwo elements of the dyad reciprocally affect themselves as a commonsystem of regulation. Thus, the lack of maternal sensitivity likely under-mines the development of emotion regulation skills in the baby.

In studying the process of development in this domain, the role ofboth internal and external sources of regulation during infancy andchildhood has been investigated. The transition from using the parentsfor regulation of arousal to being able to self-regulate is a process thatbegins early and carries on through childhood (Calkins & Fox, 2002).The caregiver's role in this process is extensive: initially, the provisionof food, clothing, and physical soothing assists the infant in state regula-tion; later, more complex communications and interactions with thecaregiver teach the child how to manage distress, control impulses,and delay gratification. The process of developing emotion regulationskills and strategies, however, is fundamentally an interactive one, anddepends on both infant and caregiver contributions. Its success or failurelikely depends on whether the goals of both participants are in agree-ment (DiSantis, Collins, Fisher, & Davey, 2011; Velotti, Zavattini, &Garofalo, 2013).

The complex interplay between mother and child is also importantin the development of the child's eating self-regulation (Anderson,Gooze, Lemeshow, & Whitaker, 2012). Indeed, because maternal sensi-tivity largely depends on the mother's ability to regulate her own emo-tions (Ammaniti, Ambruzzi, Lucarelli, Cimino, & D'Olimpio, 2004;Mills-Koonce et al., 2006), it has been suggested that the poor emotionalfunctioning of the mother might in turn lead to the establishment ofearly eating problemsof the baby, by affecting howmuch andwhat chil-dren eat during mealtime (Blissett & Farrow, 2007; Farrow & Blisset,2006; Hughes et al., 2011). For instance, if parent–child exchanges areinadequate during the feeding interactions, children may learn to usefood as a consolation instrument (Faith, Scanlon, Birch, Francis, &Sherry, 2004). Similarly, directive/controlling feeding attitudes of theparents are associated with lower self-regulation in eating and higherweight status among children (Veugelers & Fitzgerald, 2005). Alongthe same line, parents who highly control their children's food intake,lead them to pay attention to external rather than internal cues inorder to regulate the amount of food, which results in lack of self-regulation and greater eating in the absence of hunger (Hughes et al.,2011). Thus, by affecting maternal sensitivity, the maternal inability toregulate emotion likely plays a key role in the overeating behaviors ofthe child, and presumably represents an important risk factor for over-weight and obesity in developmental age (Farrow & Blisset, 2006).

1.2. Aim of the study

The link betweenmaternal emotion regulation and early eating pat-terns finds some support in the literature. However, despite the grow-ing body of research related to this topic and the increasing number ofresearchers investigating the emotional processes underlying the over-weight risk, there is still a lack of longitudinal data in this area. To over-come this lack of empirical data, the current study examined the impactof overweight andmaternal emotion dysregulation on the quality of theearly feeding interactions. Specifically, we aimed to evaluate the extentto which the quality of emotion regulation strategies in overweightpregnant women, as well as maternal overweight itself, would predictthe subsequent dysfunctional feeding interactions with their babies.

2. Method

2.1. Procedure

Prospective participants were contacted at the Isola Tiberina,Fatebenefratelli hospital, in Rome. After the approval from the hospital'sInstitutional ReviewBoardwas received, all pregnantwomenwhowerereferred for assistance to the Department of Obstetrics and Gynecologywere informed by their gynecologist about the possibility to volunteerfor the study. Women willing to participate were given a consent formand a subject's bill of rights. Confidentiality was also reviewed. Thoughnomonetary compensationwas offered in exchange for participation, atthe end of the study participants received a DVDwith the video of theirfeeding interactions.

During the beginning of the third trimester (28–40 gestationalweeks), we conducted the first screening related to the pre-pregnancyBMI. Additional medical and bio-psycho-social information was alsocollected by a gynecologist and a psychologist, so as to ascertain eligibil-ity. Participants were then assigned to a group based on their BMI: theoverweight group was made up of women with a pre-pregnancy BMIbetween 25 and 30, while the non-overweight group was made up ofwomen with a pre-pregnancy BMI between 20 and 25.

Inclusion criteria required that the participants be (a) primiparouswomenwith full-term singleton gestations, (b) not experiencing chron-ic diabetes or hypertension, (c) married or cohabiting, (d) in absence ofany full-blown psychological diagnosis, and (e) between 28 and38 years old.Women delivering preterm or post termwere also exclud-ed, in order to avoid confounds related to stressful reactions to a differ-ent medical condition. As a result of these criteria, this investigationgenerally included healthy women.

2.2. Research design

A quasi-experimental, longitudinal research design, including twoindependent groups and two research steps1, was used. As notedabove, the two groups were composed of overweight vs. non-overweight pregnant women. The first step of the research occurredduring the third trimester of pregnancy, and aimed at assessing mater-nal emotion regulation during pregnancy. The second step occurred at7-months of age of the baby, and aimed at assessing the mother–babymealtime interactions. The choice of 7 months was based on the factthat babies typically begin to eat solid foods and tomore actively partic-ipate in the feeding interactions at that age. Observing the feeding inter-actions at 7 months, thus, allowed for assessing the contribution of bothpartners, instead of only focusing on the behavior of the mother.

2.3. Participants

The entire sample was collected at the Obstetrics and GynecologyDepartment of Fatebenefratelli Isola Tiberina Hospital in Rome. At step1, the sample was comprised of 65 Italian women, 33 of whom wereoverweight and 32 non-overweight. The mean age was approximately35 years, and nearly half of the sample had a bachelor's degree or ahigher level of education. The two groups did not significantly differ interms of age, education, and gender of the baby. However, in line withother studies in the field (Veugelers & Fitzgerald, 2005), education didapproach significance when considering an alpha value of .10. A moredetailed description of the sample available at step 1 is presented inTable 1.

Of the 65 women included in the first step of the research, 12discontinued their participation after the first step, so the second stepof the research only included 53dyads. Age, education, employment po-sition and gender of the baby did not account for attrition rates. Howev-er, it should be pointed out that two-thirds of the women whodiscontinued their participation (i.e., 8 out of 12) were in the over-weight group.

Table 1Composition of the sample at step 1.

Overweight group (N = 33) Non-overweight group (N = 32)

Age (t(49) = − .80; p = .78)Mean 35.2 35.8SD 4.2 3.6

Education (phi = .20, p = .11)High school or less 22 (66.7%) 15 (46.9%)Bachelor degree or more 11 (33.3%) 17 (53.1%)

Employment (chi2(3) = 5.52, p = .14)House wife 5 (15.2%) 1 (3.1%)General employee 17 (51.5%) 13 (14.6%)Freelance 7 (21.2%) 14 (43.8%)Other 4 (12.1%) 4 (12.5%)

Gender of the baby (phi = . 14, p = .31)Male 14 (50%) 16 (64%)Female 14 (50%) 9 (36%)

405G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

2.4. Measures

During the third trimester of pregnancy (step 1) participants wereasked to complete a series of questionnaires. All were handed out per-sonally by the first author, completed at home, and then handed backto the first author. About 7 months after the delivery (step 2), themeal-time interactions between each mother and her baby were video-recorded. The questionnaires collected in step 1, as well as the feedinginteraction measure utilized in step 2, are detailed below.

2.4.1. Step 1 — Difficulties in Emotion Regulation Scale (DERS; Gratz &Roemer, 2004; Giromini, Velotti, de Campora, Bonalume, & Zavattini, 2012)

The DERS is composed of 36 items with response options on a 5-point Likert scale, ranging from 1 (almost never) to 5 (almost always).This measure assesses the following dimensions of difficulties in emo-tion regulation: lack of consciousness and understanding of emotions(awareness); nonacceptance of emotions (nonacceptance); inability tostart goal-oriented behaviors (goals); attitude toward impulsive behav-iors to face negative emotions (impulse); inaccessibility toward emo-tion regulation strategies perceived as suitable (strategies); and lack ofemotional clarity (clarity).

The Italian adaptation of the DERS was performed by Giromini et al.(2012). By investigating data from three independent Italian samples,the authors observed that the Italian DERS had good internal consisten-cy (Cronbach's alpha of .92 for the total score and an alpha≥ .77 for thesubscales), correlated significantly with a number of related constructs,and produced significantly different scores when comparing clinical vs.nonclinical adults.

In the current study, the Italian DERSwas administered during preg-nancy so as to investigate the hypothesis that the maternal ability toregulate emotion would predict the quality of the subsequent feedinginteractions with the baby.

2.4.2. Step 1 — The Symptom Checklist — 90 (SCL-90; Derogatis, 1977;Magni, Messina, De Leo, Mosconi, & Carli, 1983)

The SCL-90 is a 90-item self report symptom inventory, which pro-vides a measure of the current psychological symptom status. It isscored on nine subscales – Somatization, Obsessive–compulsive, Inter-personal Sensitivity, Depression, Anxiety, Hostility, Phobic Anxiety,Paranoid Ideation, Psychoticism – and three Global Indices of Distress(Global Severity Index, Positive SymptomDistress Index, Positive Symp-tom Total), which indicate the severity and depth of individual psycho-logical distress.

The Italian adaptation of the SCL-90 was performed by Magni et al.(1983). According to the authors, the adapted version showed satisfac-tory internal consistency (alpha ranging from .77 to .90), and high levelsof construct and convergent-discriminant validity.

Since in previous studies high levels of distress were associatedwithincreased emotional eating and eating disorders (Gilboa‐Schechtman,

Avnon, Zubery, & Jeczmien, 2006), we anticipated that the SCL-90-RGlobal Severity Index at step 1 might contribute to the predicting ofthe quality of the early feeding interactions at step 2. The Italian versionof the SCL-90 was administered so as to test this hypothesis.

2.4.3. Step 1 — Center for Epidemiological Studies — Depression Scale(CES-D; Radloff, 1977; Pierfederici et al., 1982)

The CES-D is a 20 item self-report questionnaire on depressivesymptoms. Its internal consistency is considered to be around 0.85 incommunity samples and 0.90 in psychiatric samples. Given that depres-sive symptoms often associate to eating disorders and dysfunctional dy-adic interactions (Ammaniti et al., 2004), we used the Italian version ofthe CES-D (Pierfederici et al., 1982) to assess the potential role of de-pression at step 1 in predicting the quality of the subsequent feedinginteractions.

2.4.4. Step 1—Multidimensional Scale for Perceived Social Support (MSPSS;Zimet, Dahlem, Zimet, & Farley, 1988; Prezza & Pacilli, 2002)

The MSPSS is composed of 12 items with response options on a 7-point Likert scale, ranging from 1 (absolutely false) to 7 (absolutelytrue). The instrumentmeasures support from family, friends, and signif-icant others.

The Italian adaptation of the MSPSS was carried out by Prezza andPacilli (2002). As indicated by the authors, the Italian MSPSS totalscore produced a Cronbach's alpha of .88, and a test–retest reliability(over 3 months) of r = .85.

Since social support may play a role in predicting eating behavior(Stice, 2002), we administered the Italian MSPSS so as to investigatewhether the perceived social support at step 1 might predict the subse-quent mealtime dyadic interactions.

2.4.5. Step 2 — Observational Scale for Mother–Infant Interaction duringFeeding (Chatoor et al., 1997; Lucarelli et al., 2002)

The SVIA is a 20-minute video-taped observation measuring normaland/or at-risk feeding interactions between mother and child (agerange: 1–36 months).

The Italian version of the Observational Scale for Mother–Infant In-teraction during Feeding (SVIA; Lucarelli et al., 2002) has 40 items,rated by an observer on a four-point Likert Scale, and includes four sub-scales: Affective State of themother, Interactional Conflict, Food RefusalBehaviors of the Child, and Affective State of the Dyad. Higher scores inthe Affective State of the Mother refer to greater difficulties of the care-giver to show positive affect, and a higher frequency of negative affect,such as sadness or distress. The subscale Interactional Conflict evaluatesboth the presence and intensity of exchanges of conflictwithin the dyad(e.g., the mother directs the meal according to her own emotions andintentions, rather than following the signals from the child). The sub-scale Food Refusal Behaviors of the Child explores behavioral and emo-tional characteristics of feeding patterns of the child (e.g., being easily

406 G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

distracted, opposition, and negativity). Higher scores in the AffectiveState of the Dyad refer to the difficulties of the caregiver in supportingautonomous initiatives of the child (by means of requests, insistentorders, and criticism), while the child demonstrates distress and isoppositional. As for the validity and reliability of the scale, Lucarelliet al. (2002) reported that the ICC ranged between 0.73 and 0.89(mean ICC = .83).

In the current study, the SVIAwas utilized so as tomeasure the qual-ity of themealtime interactions in the second step of the data collection,i.e., when the baby was about seven month old. All videos were inde-pendently coded by two blind reliable SVIA judges. When coding thevideos, the judges were also blind to the results of the questionnairesat step 1. Two-way random effects model single measures intraclasscorrelation coefficients (ICCs) ranged from .68 to .80, thus indicatinggood to excellent inter-rater reliability (for criteria for interpretingICCs, see Cicchetti, 1994).

2.5. Hypotheses

Based on the available literature, we hypothesized that: (a) at step 1,overweight womenwould show a higher tendency toward emotion dys-regulation, and possibly also higher psychological distress, higher depres-sion risk, and poorer perceived social support, when compared to thenon-overweight group (H1 — overweight and emotion regulationduring pregnancy); (b) dyads with overweight mothers would showhigher difficulties in feeding interactions when compared to dyads ofthe non-overweight group (H2 — overweight and feeding interactions);and (c) pre-pregnancy maternal BMI and emotion dysregulation strate-gies, and possibly also general distress, depression risk and perceivedsocial support (allmeasured during pregnancy)would predict the qualityof the feeding interactions at 7 months of the baby (H3 — predictors offeeding interactions).

3. Results

3.1. H1— overweight and emotion regulation during pregnancy

As expected, when compared to the non-overweight group, theoverweight group showed higher rates of difficulty in emotion regula-tion, with a large effect size for the total DERS score, and medium tolarge effect sizes for the subscales (Table 2). The SCL90 Global SeverityIndex also resulted in a significant difference, with a large effect size,

Table 2Differences at step 1 in emotion dysregulation, global distress, depression, and perceived socia

Overweight group(N = 33)

Non-overweightgroup (N = 32)

M SD M

DERSNonacceptance 13.2 5.5 10.4Goals 13.1 5.2 10.9Impulse 12.3 4.5 8.8Awareness 14.1 3.6 12.8Strategies 14.6 5.3 11.1Clarity 8.1 2.4 7.3Total 75.4 17.5 61.3

SCL-90Global Severity Index 0.7 0.5 0.4

CES-DTotal score 15.1 8.4 11.8

MSPSSb

Significant others 25.5 2.5 26.1Family 22.9 6.3 24.4Friends 22.3 4.7 23.5Total score 70.8 8.3 74.0

a Because homoscedasticity could not be assumed, Welch–Satterthwaite method was used tb One record in the overweight group was missing MSPSS information.

while the depression scale (CES-D) yielded only marginally significantresults (although in the expected direction), and there were no signifi-cant differences in terms of perceived social support (MSPSS). Taken to-gether, these results indicate that maternal pre-pregnancy overweightis associated, during pregnancy, with difficulties in emotion regulationand psychological distress.

3.2. H2 — overweight and feeding interactions

As shown in Table 3, the quality of the dyadic feeding interactions at7 months of age of the baby was significantly different between theoverweight and the non-overweight groups. More precisely, whencompared to the non-overweight group, the overweight group demon-strated poorer quality of feeding interactions on all the dimensions ofthe SVIA. The effect size of these differences was large for AffectiveState of the Mother, Interactional Conflict, and Affective State of theDyad, and medium for Food Refusal Behavior of the Child. Thus, as ex-pected, the overweight group was more distressed and emotionallydysregulated than the non-overweight group during pregnancy, andshowed poorer quality of feeding interactions at 7 months of age ofthe baby.

3.3. H3 — predictors of feeding interactions

A third aim of the study was to investigate whether pre-pregnancymaternal BMI, as well as maternal difficulties in emotion regulation,would predict the quality of the feeding interactions within the dyadsat 7 months of age of the baby. The potential predictive role of psycho-logical distress, depression, and perceived social support during preg-nancy was also investigated. As shown in Table 4, a large amount ofsignificant correlations were obtained. The pre-pregnancy maternalBMI, the total DERS score, and the SCL90 Global Severity Index signifi-cantly correlatedwith all 4 subscales of the SVIA, the CES-D significantlycorrelatedwith 3 SVIA subscales, and the totalMSPSS score significantlycorrelated with the Affective State of the Dyad subscale of the SVIA.

To further investigate the relationship of maternal pre-pregnancyBMI, difficulties in emotion regulation, psychological distress, depres-sion, and perceived social support to the quality of feeding interactionsat 7 months of age of the baby, a series of multiple regressions weretested. For each model, a stepwise selection method was used, inorder to identify the best predictors. The pre-pregnancy BMI, DERS, SCL-90, CES-D, and MSPSS scores were used as predictors, and each SVIA

l support between overweight and non-overweight groups.

t df p d

SD

3.7 2.33 55.8a 0.02 0.583.3 2.05 54.7a 0.05 0.512.0 4.07 44.2a 0.00 1.003.0 1.46 63 0.15 0.362.9 3.34 49.5a b0.01 0.821.6 1.65 63 0.10 0.418.3 4.16 46.2a b0.01 1.02

0.2 3.75 47.4a b0.01 0.92

6.6 1.76 63 0.08 0.44

2.7 −0.87 62 0.39 −0.224.0 −1.14 52.6a 0.26 −0.283.8 −1.08 62 0.28 −0.277.8 −1.59 62 0.12 −0.40

o adjust degrees of freedom.

Table 3Differences in feeding interactions between overweight and non-overweight groups.

Overweight group(N = 25)

Non-overweightgroup (N = 28)

t df P d

M SD M SD

Feeding interaction scaleAffective State of the Mother 11.6 3.4 8.6 3.4 3.25 51.0 b0.01 0.89Interactional Conflict 18.3 5.9 12.7 4.2 4.02 51.0 b0.01 1.11Food Refusal Behavior 10.0 2.8 8.5 2.1 2.32 51.0 0.02 0.64Affective State of the Dyad 5.0 2.4 2.9 1.3 3.96 36.5a b0.01 1.12

a Because homoscedasticity could not be assumed, Welch–Satterthwaite method was used to adjust degrees of freedom.

407G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

subscale as the dependent variable. In order to avoid multicollinearity,only the total scores were used for the DERS and MSPSS. The resultingmultiple regression models are presented in Table 5. The total DERSscore and the pre-pregnancy maternal BMI were the best predictors forAffective State of the Mother and for the Interactional Conflict, the totalDERS score was the best predictor for Food Refusal Behavior of theChild; the pre-pregnancy maternal BMI was the best predictor for Affec-tive State of the Dyad. Thus, the DERS measured during pregnancy,alongwith the pre-pregnancy BMI,were the best predictors of the qualityof the dyadic feeding interactions at 7 months of age of the baby.

4. Discussion

Some empirical data indicate that the risk for childhood obesity andoverweight increases dramatically when one or both parents are over-weight or obese. However, it has also been shown that such an associa-tion cannot be entirely explained only by biological factors (Kral & Faith,2007). Based on the available literature, we hypothesized that, in addi-tion to a number of other non-biological risk factors (Owen et al., 2005),the maternal ability to regulate emotions might also play an importantrole in explaining the intergenerational transfer of overweight and obe-sity. Specifically, we hypothesized that maternal emotion regulationwould impact the quality of the feeding interactions between themoth-er and the baby, which in turn is known as an important predictor ofchildhood obesity and overweight (Rising & Lifshitz, 2005).

By adopting a quasi-experimental, longitudinal research design, weshowed that, when compared to the non-overweight group, the over-weight group had more difficulties in emotion regulation, was more

Table 4Correlation between maternal psychological features during pregnancy and quality of feeding

SVIA SVIA

Aff. State Mother Interac

BMIMaternal pre-pregnancy BMI 0.42⁎⁎ 0.55⁎

DERSNonacceptance 0.27⁎ 0.34⁎

Goals 0.34⁎ 0.25Impulse 0.41⁎⁎ 0.36⁎

Awareness −0.11 0.04Strategies 0.49⁎⁎ 0.51⁎

Clarity 0.25 0.11Total 0.43⁎⁎ 0.44⁎

SCL-90Global Severity Index 0.36⁎⁎ 0.38⁎

CES-DTotal score 0.32⁎ 0.30⁎

MSPSSa

Significant Others −0.13 −0.20Family −0.17 −0.04Friends −0.06 −0.16Total score −0.18 −0.17

a One record in the overweight group was missing MSPSS information.⁎ p b 0.05.⁎⁎ p b 0.01.

psychologically distressed, and had poorer feeding interactions withtheir babies. Perhaps more importantly, the extent to which theparticipants were suffering difficulties in emotion regulation duringpregnancy predicted, significantly and beyond the effects of pre-pregnancy BMI, the quality of the subsequent dyadic feeding interac-tions, 7 months after the delivery. To the best of our knowledge, thisis the first study to investigate the predictive role of maternal emotionregulation during pregnancy on the subsequent feeding interaction.

According to our first hypothesis (H1 — overweight and emotionregulation during pregnancy) we predicted that overweight motherswould be more likely than non-overweight mothers to show emotionregulation difficulties, during pregnancy. As shown in Table 2, this hy-pothesis was fully supported by our results. In particular, not only didthe overweight group show higher difficulties in emotion regulation,but it also showed higher levels of psychological distress and (albeitonly marginally significantly) higher depression risk. Given the amountof research indicating that eating disorders associate to deficits in thecognitive processing of emotions (de Groot, Rodin, & Olmsted, 1995),this finding is perhaps not surprising. Indeed, our suggestion is thatemotion dysregulation and disordered eating are so intrinsically relatedto each other, that the ability to regulate the emotion should be assessedevery time an existing overweight condition cannot be explained bybiological factors.

Our second hypothesis (H2 — overweight and feeding interactions)anticipated that the overweight group would show more problematicdyadic mealtime interactions, when compared to the non-overweightgroup. This hypothesis was based on the idea that the overweight con-dition is thought to be associated with emotion and eating regulation

interaction at 7 months of age (N = 53).

SVIA SVIA

tional Conflict Food Refusal Behavior Aff. State Dyad

⁎ 0.28⁎ 0.55⁎⁎

0.32⁎ 0.30⁎

0.34⁎ 0.11⁎ 0.32⁎ 0.28⁎

−0.07 −0.11⁎ 0.45⁎⁎ 0.40⁎⁎

−0.05 0.01⁎ 0.38⁎⁎ 0.29⁎

⁎ 0.32⁎ 0.34⁎

0.30⁎ 0.25

−0.10 −0.32⁎

−0.12 −0.08−0.10 −0.25−0.16 −0.29⁎

Table 5Multiple regressionmodelswith pre-pregnancy BMI, DERS, SCL90Global Severity Index, CES-D, andMSPSS as predictors (stepwisemethod) and each SVIA subscale as criterion (N = 52).

Criterion/predictors entered by step β1 β2 R R2 Adj. R2 ΔR2

SVIA Affective State of the MotherStep 1 0.43 0.19 0.17 –

Total DERS score 0.43⁎⁎ 0.32⁎

Step 2 0.52 0.27 0.24 0.08⁎

Pre-pregnancy BMI – 0.31⁎

SVIA Interactional ConflictStep 1 0.55 0.31 0.29 –

Pre-pregnancy BMI 0.55⁎⁎ 0.46⁎⁎

Step 2 0.61 0.37 0.34 0.06⁎

Total DERS score – 0.27⁎

SVIA Food Refusal BehaviorStep 1 0.38 0.15 0.13 –

Total DERS score 0.38⁎⁎ –

SVIA Affective State of the DyadStep 1 0.55 0.30 0.29 –

Pre-Pregnancy BMI 0.55⁎⁎ –

β1, β2 = standardized beta coefficients for steps 1 and 2.⁎ p b 0.05.⁎⁎ p b 0.01.

408 G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

issues (Sim& Zeman, 2005;Whiteside et al., 2007), and that therefore itshould also be characterized by poor maternal sensitivity (Andersonet al., 2012). As shown in Table 3, this hypothesis was also fully con-firmed. In fact, when compared to the non-overweight group, the over-weight mothers: (a) showed less positive affect (SVIA— Affective Stateof the Mother); (b) were more prone to direct the meals according totheir own emotions and intentions, rather than following the signalsfrom the child (SVIA— Interactional Conflict); (c) producedmore dyad-ic exchanges characterized by opposition, distraction, and negativity ofthe baby (SVIA — Food Refusal Behavior of the Child); and (d) showedgreater difficulties in supporting autonomous initiatives of the baby(SVIA— Affective State of the Dyad). These findings are particularly im-portant in terms of understanding the early risk factors for overweightand obesity. Indeed, a lack of maternal sensitivity during the mealtimeinteractions (e.g., lack of sensitivity for the child's choices and prefer-ences, as well as for the child's emotional state) is thought to reducethe baby's competence to employ its own hunger and satiety cues,which in turn represents an important risk factor for future eating disor-ders (Rising & Lifshitz, 2005).

The third aim of the current study (H3 — predictors of feedinginteractions) was to investigate whether the quality of emotion regula-tion strategies measured during pregnancy, as well as the maternalpre-pregnancy BMI, would predict the subsequent feeding interactionsof the dyads. The results presented in Tables 4 and 5 show that themostimportant contributors to the prediction of the SVIA scores (measuredat the second step of the research) were the pre-pregnancy BMI andthe total DERS scores (both measured during the first step of the re-search). In fact, when testing a series of multiple regressions with astepwise method, part of the variance of the SVIA scores appeared tobe uniquely explained by the maternal pre-pregnancy BMI, and part ofit appeared to be uniquely explained by thematernal ability to regulatethe emotion. Thus, in line with other studies indicating that the inter-generational transfer of overweight and obesity cannot be ascribedonly to biological factors (Agras, Hammer, McNicholas, & Kraemer,2004; Kral & Faith, 2007), these findings provide evidence that the ma-ternal ability to regulate emotions, as measured during pregnancy, maypredict the quality of the mealtime interactions seven months after thedelivery.

Many researchers have supported the idea that emotion regulationproblems work as a maintenance factor for eating problems, and thateating problems work as a means of regulating negative affect (Svaldi,Griepenstroh, Tuschen-Caffier, & Ehring, 2012). Our suggestion is thatmaternal emotion dysregulation might promote and maintain the

overweight of the mothers, and also be transmitted within the feedinginteraction with the baby. As noted above, indeed, a low maternal sen-sitivity for the child's own choices and preferences, as well as for his orher emotional states, likely affects the child's ability to develop self-regulation skills, and poses him or her at risk for learning to self-regulate himself or herself through food intake. If themealtime interac-tions lack maternal sensitivity, indeed, the children may learn to usefood as a consolation instrument, ormay learn to pay attention to exter-nal rather than internal cues in order to regulate the amount of food(Hughes et al., 2011). This, in turn,may constitute a risk factor for eating(Frankel et al., 2012), and perhaps also for emotion dysregulation.

To date, the great majority of the studies on these topics haveadopted cross-sectional research designs, which do not allow to appro-priately investigate causal relationships. In contrast, by adopting a longi-tudinal research design, the current study has probably provided moreconclusive evidence for the existence of a causal relationship betweencertainmaternal characteristics and their subsequent dyadic feeding in-teractions. Nevertheless, it should be noted that a number of limitationsalso characterize ourwork. First of all, the ability to regulate the emotion– a key variable of this investigation – was only measured through aself-report instrument. Although the DERS is fairly consolidated in theliterature as a well validated instrument, social desirability and otherpotential biases might still have occurred. Second, while we aimed atidentifying early risk indicators associated with the overweight condi-tion, given the age of the babies we were not able to provide actual out-come measures for the weight of the babies. Future follow-ups mightprovide important information, in regard to this aspect. Third, the sam-ple size was relatively small, and some attrition occurred, so that someof the analyses lacked adequate power to support our conclusions.Given that about 18% of the participants did not complete the study,in particular, it is difficult to draw precise conclusions in regard to thegeneralizability of our findings. The fact that none of the demographicvariables under investigation significantly associated with the attritionrates, however, suggests that attrition did not act as a confound, inthis study. Fourth, our data do not refer to breastfeeding or the bottlefeeding, while research indicates that being bottle fed may increasethe risk of lower self regulation skills of the baby. Future research shouldconsider including this measure.

With these limitations in mind, the current study is the first tosuggest that emotion dysregulation during pregnancy can be pre-dictive of the subsequent development of feeding interaction prob-lems, and paves the way for future follow-ups and replicationstudies.

409G. de Campora et al. / Eating Behaviors 15 (2014) 403–409

Role of funding sourcesThis research is part of a dissertation study andno funding for this studywasprovided.

ContributorsAuthor A designed the study, and wrote the protocol. Author B conducted the statis-

tical analyses. Authors A and Bwrote the first draft of themanuscript. Author C supervisedthe inclusion criteria, selected the eligible participants, and provided feedback acrossvarious draft. Author D conducted literature searches and provided summaries of previousinvestigations. Author E supervised the research, and provided feedback across the datacollection. All authors contributed to and have approved the final manuscript.

Conflict of interestsNone of the authors have any conflict of interest to declare.

AcknowledgmentsWe thank to Dr. Elio Cirese, M.D., Head of the OBGYN Department, for his availability

and support at the beginning of this research; the Department of Dynamic and ClinicalPsychology and the Lab of Couple and Family relationships of Sapienza University ofRome, for their help and sustain throughout the research; Verdiana Imperio and VanessaPalombi, for their help with data collection; and Dr. Cher Rafiee for proofreading themanuscript.

References

Agras, W. S., Hammer, L. D., McNicholas, F., & Kraemer, H. C. (2004). Risk factors forchildhood overweight: A prospective study from birth to 9.5 years. Journal ofPediatrics, 145, 20–25.

Agras, W. S., & Mascola, A. J. (2005). Risk factors for childhood obesity. Current Opinion inPediatrics, 17, 648–652.

Ammaniti, M., Ambruzzi, A. M., Lucarelli, L., Cimino, S., & D'Olimpio, F. (2004). Malnutri-tion and Dysfunctional Mother–Child Feeding Interactions: Clinical assessment andresearch implications. Journal of the American College of Nutrition, 23(3), 259–271.

Anderson, S. E., Gooze, R. A., Lemeshow, S., & Whitaker, R. C. (2012). Quality of earlymaternal–child relationship and risk of adolescent obesity. Pediatrics, 129(1),132–140.

Blissett, J., & Farrow, C. (2007). Predictors of maternal control of feeding at 1 and 2 yearsof age. International Journal of Obesity, 31(10), 1520–1526. http://dx.doi.org/10.1038/sj.ijo.0803661.

Calkins, S. D., & Fox, N. A. (2002). Self-regulatory processes in early personality develop-ment: A multilevel approach to the study of childhood social withdrawal and aggres-sion. Development and Psychopathology, 14, 477–498. http://dx.doi.org/10.1017/S095457940200305X.

Chatoor, I., Getson, P., Menvielle, E., Brasseaux, C., O'Donnell, R., Rivera, Y., et al. (1997). Afeeding scale for research and clinical practice to assess mother–infant interactions inthe first three years of life. Infant Mental Health Journal, 18(1), 76–91.

Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed andstandardized assessment instruments in psychology. Psychological Assessment, 6(4),284.

de Groot, J. M., Rodin, G., & Olmsted, M. P. (1995). Alexithymia, depression, and treatmentoutcome in bulimia nervosa. Comprehensive Psychiatry, 36(1), 53–60.

Derogatis, L. R. (1977). SCL-90. Administration, scoring & procedures manual — I for ther(evised) version and other instruments of the psychopathology rating scale series.Baltimore: John Hopkins University Press.

DiSantis, K. I., Collins, B. D., Fisher, J. O., & Davey, A. (2011). Do infants fed directly from thebreast have improved appetite regulation and slower growth during early childhoodcompared with infants fed from a bottle? International Journal of Behavioral Nutritionand Physical Activity, 8, 89.

Faith, M. S., Scanlon, K. S., Birch, L. L., Francis, L. A., & Sherry, B. (2004). Parent–childfeeding strategies and their relationships to child eating and weight status. ObesityResearch, 12, 1711–1722. http://dx.doi.org/10.1038/oby.2004.212.

Farrow, C., & Blisset, J. (2006). Maternal cognitions, psychopathologic symptoms, andinfant temperament as predictors of early infant feeding problems: A longitudinalstudy. International Journal of Eating Disorders, 39(2), 128–134. http://dx.doi.org/10.1002/eat.20220.

Farrow, C. V., & Blissett, J. (2008). Controlling feeding practices: Cause or consequence ofearly child weight? Pediatrics, 121(1), e164–e169. http://dx.doi.org/10.1542/peds.2006-3437.

Frankel, L. A., Hughes, S. O., O'Connor, T. M., Power, T. G., Fisher, J. O., & Hazen, N. L. (2012).Parental influences on children's self-regulation of energy intake: Insights fromdevelopmental literature on emotion regulation. Journal of Obesity. http://dx.doi.org/10.1155/2012/327259.

Gilboa‐Schechtman, E., Avnon, L., Zubery, E., & Jeczmien, P. (2006). Emotional processingin eating disorders: Specific impairment or general distress related deficiency?Depression and Anxiety, 23(6), 331–339.

Giromini, L., Velotti, P., de Campora, G., Bonalume, L., & Zavattini, G. C. (2012). Culturaladaptation of the difficulties in emotion regulation scale: Reliability and validity ofan Italian version. Journal of Clinical Psychology, 00, 1–20. http://dx.doi.org/10.1002/jclp.21876.

Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of emotion regulation anddysregulation: Development, factor structure, and initial validation of the difficultiesin emotion regulation scale. Journal of Psychopathology and Behavioral Assessment,26(1), 41–54. http://dx.doi.org/10.1007/s10862-008-9102-4.

Hughes, S. O., Power, T. G., Papaioannou, M. A., Cross, M. B., Nicklas, T. A., Hall, S. K., et al.(2011). Emotional climate, feeding practices, and feeding styles: An observationalanalysis of the dinner meal in head start families. International Journal of BehavioralNutrition and Physical Activity, 8(60), 1–11. http://dx.doi.org/10.1186/1479-5868-8-60.

Kral, T. V., & Faith, M. S. (2007). Child eating patterns and weight regulation: A develop-mental behaviour genetics framework. Acta Paediatrica, 96(s454), 29–34.

Lucarelli, L., Cimino, S., Perucchini, P., Speranza, A. M., Ammaniti, M., & Ercolani, A. P.(2002). I disturbi alimentari nella prima infanzia: validazione di uno strumentoosservativo dell'interazione madre-bambino. Infanzia e Adolescenza, 2, 113–124.

Magni, G., Messina, C., De Leo, D., Mosconi, A., & Carli, M. (1983). Psychological distress inparents of children with acute lymphatic leukemia. Acta Psychiatrica Scandinavica, 68,297–300.

Mills-Koonce, W. R., Gariépy, J. L., Propper, C., Sutton, K., Calkins, S., Moore, G., et al.(2006). Infant and parent factors associated with early maternal sensitivity: Acaregiver-attachment systems approach. Infant Behavior and Development, 30(1),114–126. http://dx.doi.org/10.1016/j.infbeh.2006.11.010.

Owen, C. G., Martin, R. M., Whincup, P. H., Smith, G. D., & Cook, D. G. (2005). Effect ofinfant feeding on the risk of obesity across the life course: A quantitative review ofpublished evidence. Pediatrics, 115(5), 1367–1377. http://dx.doi.org/10.1542/peds.2004-1176.

Pierfederici, A., Fava, G. A., Munari, F., Rossi, N., Baldaro, B., Grandi, S., et al. (1982).Validazione italiana del CES-D per la misurazione della depressione. In R. Canestrari(Ed.), Nuovi Metodi in Psicometria (pp. 95–103). Firenze: Organizzazioni Speciali.

Prezza, M., & Pacilli, M. G. (2002). Perceived social support from significant others, familyand friends and several socio-demographic characteristics. Journal of Community &Applied Social Psychology, 12(6), 422–429. http://dx.doi.org/10.1002/casp.696.

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in thegeneral population. Applied Psychological Measurement, 1(3), 385–401. http://dx.doi.org/10.1177/014662167700100306.

Rising, R., & Lifshitz, F. (2005). Relationship betweenmaternal obesity and infant feeding-interactions. Nutrition Journal, 4(17.2005).

Schore, A. N. (2000). Attachment and the regulation of the right brain. Attachment &Human Development, 2(1), 23–47. http://dx.doi.org/10.1080/146167300361309.

Sim, L., & Zeman, J. (2005). Emotion regulation factors as mediators between bodydissatisfaction and bulimic symptoms in early adolescent girls. The Journal of EarlyAdolescence, 25(4), 478–496. http://dx.doi.org/10.1177/0272431605279838.

Stice, E. (2002). Risk and maintenance factors for eating pathology: A meta-analyticreview. Psychological Bulletin, 128, 825–848.

Svaldi, J., Griepenstroh, J., Tuschen-Caffier, B., & Ehring, T. (2012). Emotion regulation def-icits in eating disorders: A marker of eating pathology or general psychopathology?Psychiatry Research, 197(1), 103–111.

Van der Merwe, M. T. (2007). Psychological correlates of obesity in women. InternationalJournal of Obesity, 31, S14–S18. http://dx.doi.org/10.1038/sj.ijo.0803731.

Velotti, P., Zavattini, G. C., & Garofalo, C. (2013). Lo studio della regolazione delleemozioni: Prospettive future. Giornale Italiano di Psicologia, 2, 247–262.

Ventura, A., & Birch, L. (2008). Does parenting affect children's eating and weight status?International Journal of Behavioral Nutrition and Physical Activity, 5(1), 15. http://dx.doi.org/10.1186/1479-5868-5-15.

Veugelers, P. J., & Fitzgerald, A. L. (2005). Prevalence of and risk factors for childhoodoverweight and obesity. Canadian Medical Association Journal, 173(6), 607–613.

Whiteside, U., Chen, E., Neighbors, C., Hunter, D., Lo, T., & Larimer, M. (2007). Difficultiesregulating emotions: Do binge eaters have fewer strategies to modulate and toleratenegative affect? Eating Behaviors, 8(2), 162–169.

Zimet, G. D., Dahlem, N.W., Zimet, S. G., & Farley, G. K. (1988). Themultidimensional scaleof perceived social support. Journal of Personality Assessment, 52(1), 30–41.