Article Text
PDF
Abstract
Objective To examine the association between parental anxiety and depression with child fussy eating—that is, consistent rejection of particular food items.
Design This study was embedded in Generation R, a prospective cohort from fetal life onwards in the Netherlands.
Setting Population-based.
Participants 4746 4-year-old children and their parents.
Exposure Parental internalising problems (ie, symptoms of anxiety and depression) were assessed with the Brief Symptoms Inventory during pregnancy and the preschool period (child age 3 years).
Main outcome measure The food fussiness scale of the Children's Eating Behaviour Questionnaire.
Results Maternal anxiety during pregnancy and during the child's preschool period was related to higher food fussiness sum-scores in children. For instance, per point on the anxiety scale in pregnancy, children had on average a 1.02 higher sum-score (95% CI 0.59 to 1.46) on the food fussiness scale, after adjustment for confounders. Likewise, mothers’ depressive symptoms at both time points were associated with fussy eating behaviour in their children (eg, in the antenatal period: per point on the depression scale, children had a 0.91 point higher sum-score on the food fussiness scale, 95% CI 0.49 to 1.33). We found largely similar associations between fathers’ internalising problems and children's fussy eating. However, fathers’ anxiety during the antenatal period was not related to child fussy eating.
Conclusions Maternal and paternal internalising problems were prospectively associated with fussy eating in preschoolers. Healthcare practitioners should be aware that non-clinical symptoms of anxiety and depression in parents are risk factors for child fussy eating.
- Child Psychology
- Epidemiology
- Nutrition
Statistics from Altmetric.com
What is already known on this topic
Fussy eating behaviour, characterised by consistently rejecting particular food items, is a common problem in childhood causing major concerns among parents.
Maternal postnatal anxiety and depressive symptoms have been related to child fussy eating.
It is unclear whether maternal internalising problems are risk factors for child fussy eaters or rather a result of child fussy eating behaviour.
What this study adds
Mothers’ antenatal anxiety symptoms predict child fussy eating independent of mothers’ postnatal anxiety symptoms.
Also mothers’ depressive symptoms during pregnancy and 3 years later were associated with more fussy eating in their children.
We found indications that fathers’ internalising problems are related to children's fussy eating.
Introduction
Fussy eating is characterised by consistent rejection of particular foods, which results in a restricted diet variety,1 causing major concerns among parents.2 Child fussy eating has been associated with functional constipation,3 weight problems,4 and behavioural problems.5 Previous research suggested parental controlling feeding4 and parental physical and mental health problems6 as potential risk factors for fussy eating (also called ‘picky’ or ‘selective’ eating). However, the aetiology of fussy eating is not well understood.6
It is well known that internalising psychiatric problems of parents (ie, anxiety and depression) are related to problematic child development,7 including disturbed eating behaviours.8 ,9 A complex interplay of multiple factors such as genetics, disturbed parent–child interaction, and modelling of parent behaviour account for the increased risk of problems in children,7 and may also affect children's fussy eating. Maternal internalising problems have been related to fussy eating in preschool-aged children in population-based studies.6 ,10 Maternal internalising problems during the child's preschool period have been found to be predictive of persistent fussy eating at a later age.11 Farrow and Blissett,12 however, have reported that antenatal and postnatal maternal psychiatric symptoms did not predict fussy eating in 6-month-old children.
Most studies have focused on symptoms during the child's preschool period. This is a sensitive period in development, but the Barker hypothesis13 highlights the need to also study antenatal anxiety and depression. Another advantage of studying internalising problems in the antenatal period is that the association with children's fussy eating is less prone to reverse causation—that is, children's fussy eating is not likely to affect their mothers’ problems during pregnancy. In addition, most previous studies were limited in their reliance on maternal reports of both exposure (internalising problems) and outcome (child fussy eating).6 ,10 ,11 ,14 Consequently, reported associations may be overestimated because of reporter bias, as the depression-distortion hypothesis states that mothers with psychiatric problems might have a biased perception of their child's behaviour.15 Last, most studies have focused on mothers’ anxiety and depression without studying the effects of fathers’ symptoms.
The present study's objective was to examine whether maternal and paternal internalising problems are prospectively associated with children's fussy eating, using multiple informants of child eating behaviour. More specifically, we aimed to evaluate the role of anxiety and depressive symptoms in the antenatal and preschool period.
Materials and methods
Study design and procedure
This study was embedded in Generation R, a population-based prospective cohort from fetal life onwards.16 ,17 Pregnant women living in Rotterdam, The Netherlands, with a delivery date between April 2002 and January 2006 were invited to participate (response rate 61%). The local medical ethics committee approved the study. Sociodemographic information was collected by postal questionnaire during pregnancy and from medical birth records completed by gynaecologists and midwives. Parental internalising problems were assessed by postal questionnaire during mid-pregnancy, and again when the child was 3 years old. At 3 and 4 years of age, parents filled in postal questionnaires including an assessment of their children's eating behaviour. More detailed information about the design and procedure is available elsewhere.16
Participants
Parents of 7295 children gave full consent for the preschool phase of Generation R. Those with missing data on the food fussiness scale of the Children's Eating Behaviour Questionnaire (CEBQ) were excluded (N=2355). Of the remaining parent–child dyads, 194 participants had missing values for maternal anxiety or depression during pregnancy and 3 years later, yielding a sample size of 4746 children and mothers. The population for analysis with fathers’ anxiety or depression was smaller (N=4144), as 602 participants had missing values for fathers’ anxiety or depression on both time points.
Measures
Parental anxiety and depressive symptoms
Anxiety and depressive symptoms of both mothers and fathers were assessed with the Brief Symptom Inventory (BSI) at two time points: during mid-pregnancy and 3 years later. The BSI is a validated 53-item self-report questionnaire assessing a spectrum of psychological problems in the preceding 7 days.18 ,19 We used the anxiety scale (eg, ‘feeling fearful’) and the depression scale (eg, ‘feeling lonely’). Each scale consists of six items rated on a 5-point Likert scale from 0 (not at all) to 4 (extremely). For each scale, mean scores were calculated, with higher scores indicating more problems.
Child fussy eating behaviour
At age 4 years, fussy eating was assessed with the CEBQ, a validated parent report questionnaire.20 The CEBQ consists of eight subscales, containing 35 items on which parents rate the frequency of their children's eating behaviours. We used the subscale food fussiness, which consists of six items covering children who are difficult to please with meals, who display food neophobia (eg, ‘My child refuses new foods at first’), and who have a limited diet variety (eg, ‘My child enjoys a wide variety of foods’, reverse coded). Each item was answered on a Likert-type scale from 1 (never) to 5 (always). Scale sum-scores were calculated, with higher scores indicating more food fussiness (range 6–30). To facilitate comparison with other studies, we also calculated the mean score of this scale.
As most CEBQs were filled out by mothers (∼88%), we also used the Child Behaviour Checklist for toddlers (CBCL/1½–5)21 for which we had multiple informants. Two items were used as a proxy for fussy eating: (1) ‘does not eat well’ and (2) ‘refuses to eat’ in the past 2 months. These questions were answered by both mothers and fathers when the children were 3 years old. Items were rated on a 3-point Likert scale from 1 (not true) to 3 (often true). Sum-scores were calculated (range 2–6) and children with a score of ≥4 were classified as ‘fussy eaters’.22
Confounders
During pregnancy, questionnaires were used to assess sociodemographic characteristics: parental age, family income, parental ethnic background (based on country of birth of parents and grandparents), parental educational level, marital status, and parity (defined as number of live births mothers delivered before birth of the participating child). Mode of delivery, sex of child, and birth characteristics (birth weight and gestational age) were obtained from medical records.
Statistical analysis
We used separate linear regression analyses to test whether higher scores of mothers’ anxiety and depression on the BSI at each time point (during pregnancy and at 3 years postnatal) were related to higher sum-scores on the CEBQ's food fussiness scale. We also tested the independent effects of maternal internalising problems on child food fussiness by analysing the two time points in the same model. In addition, we explored whether fathers’ anxiety and depression scores at each time point were related to food fussiness, using separate linear regression analyses. All antenatal models were adjusted for sociodemographic characteristics. All models with postnatal internalising symptoms were additionally adjusted for mode of delivery, sex of child, and birth characteristics.
We performed several sensitivity analyses. As the CEBQ's food fussiness scale was mainly reported by mothers, the associations of maternal anxiety and depression with this outcome measure may be prone to reporter bias. Therefore, we additionally examined the associations of mothers’ anxiety and depression scales (continuously) with the CBCL data on fussy eating, as obtained by multiple informants. Separate logistic regression analyses were conducted for mother reports and father reports on the CBCL. Second, using linear regression analyses, we compared the food fussiness scores of the following three groups of children: (1) children of mothers who had average or below average anxiety or depression scores (reference group); (2) children of mothers who had above average anxiety scores (0.50 and higher but below clinical cut-off) or above average depression scores (0.33 and higher but below clinical cut-off); (3) children of mothers who had clinically significant anxiety scores (0.71 and higher) or clinically significant depression scores (0.80 and higher). The Dutch cut-offs for the BSI were used to categorise the mothers.23 All sensitivity analyses were adjusted for the same potential confounders as described above.
Multiple imputation techniques were used to impute missing values on confounders and exposure.24 The reported B-values are pooled from 20 imputed datasets. In addition, we repeated our main analyses in complete cases. All statistical analyses were performed with SPSS V.21.0.
Results
Sample characteristics
Sample characteristics are presented in table 1. The food fussiness sum-score at age 4 was 17.7 (SD=4.9), and the mean was 3.0 (SD=0.8). Using the CBCL as proxy for fussy eating, ∼30% of all children were classified as fussy eaters at age 3. In total, agreement of mothers and fathers about their child being a fussy or non-fussy eater was 76.7%. We calculated Yule's Y25 to be 0.47, indicating moderate agreement between mothers and fathers.
Sample characteristics of 4746 parent–child dyads in the Generation R Study
Parental anxiety symptoms and children's fussy eating behaviour
Maternal anxiety symptoms during pregnancy and during the preschool period were related to fussy eating in their 4-year-old children (table 2). For instance, per point on the anxiety scale in pregnancy, children had a 1.02 higher food fussiness sum-score (95% CI 0.59 to 1.46). By additionally analysing maternal anxiety at both time points in the same model, we found that mothers’ anxiety during pregnancy (B=0.81, 95% CI 0.33 to 1.29) and during the preschool period (B=0.54, 95% CI 0.05 to 1.03) were both independently related to child fussy eating (not shown in tables). Fathers' anxiety in the preschool period, but not during the antenatal period, was related to fussy eating in their child (table 2).
Parental anxiety symptoms and fussy eating in 4-year-old children (CEBQ)
Sensitivity analyses showed that not only children of mothers with clinically significant anxiety had elevated food fussiness scores (eg, antenatal model: B=1.06, 95% CI 0.46 to 1.67), but children of mothers with anxiety scores above average also had higher food fussiness scores than children of mothers with average or below average anxiety scores (eg, antenatal model: B=0.72, 95% CI 0.21 to 1.22) (see online supplementary table S1).
Parental depressive symptoms and children's fussy eating behaviour
Table 3 shows that higher maternal depressive symptoms in the antenatal period as well as at 3 years postnatal were related to more fussy eating in their 4-year-old children (eg, per point antenatal depression score, children had a 0.91 higher food fussiness sum-score, 95% CI 0.49 to 1.33). Likewise, the associations between fathers’ depressive symptoms at both time points and children's food fussiness were in the same direction (table 3).
Parental depressive symptoms and fussy eating in 4-year-old children (CEBQ)
Similar to the independent effects of mothers’ anxiety at both time points, we also found that mothers’ depressive symptoms during pregnancy and 3 years later were independently related to child fussy eating (data not shown). Online supplementary table S2 shows that mothers’ depression scores above average already predicted fussy eating, especially during pregnancy (B=0.87, 95% CI 0.41 to 1.33 for above average scores and B=0.87, 95% CI 0.22 to 1.51 for clinically significant depression).
Mothers’ internalising problems and children's fussy eating across informants
Table 4 shows that maternal internalising problems were also associated with both mother and father reports of children's fussy eating on the CBCL. ORs were very similar regardless of whether mothers or fathers reported their 3-year-olds’ fussy eating behaviour (eg, for antenatal anxiety OR=1.50 (95% CI 1.18 to 1.89) as reported by mothers and OR=1.44 (95% CI 1.13 to 1.83) as reported by fathers).
Maternal internalising problems and a proxy for fussy eating at age 3 years as independently reported by both parents on the CBCL
Additional sensitivity analyses
Results of our full case analyses (see online supplementary tables 3–5) were very similar to our main findings (tables 2 and 3). For instance, mothers’ internalising problems were also related to child food fussiness at age 4 (eg, for antenatal anxiety B=1.15 (95% CI 0.69 to 1.61)). Only the associations of fathers’ internalising problems in the preschool period with fussy eating were no longer statistically significant, probably because of reduced power, although the magnitude of the associations was also slightly reduced (eg, for anxiety B=0.69 (95% CI −0.14 to 1.52)).
Discussion
Higher maternal internalising problems during pregnancy and at 3 years postnatal were prospectively and both independently related to child fussy eating in a large population-based cohort. We also found indications that paternal internalising problems are related to child fussy eating.
The finding that maternal internalising problems predicted more fussy eating in children is largely consistent with previous research,6 ,9–11 ,26–28 although conflicting studies exist.12 ,29 Importantly, we found that mothers’ antenatal internalising symptoms predicted 4-year-olds’ fussy eating independent of mothers’ symptoms at 3 years postnatal. This strongly suggests that the direction of the associations with mothers’ antenatal symptoms is from mother to child. Coulthard and Harris30 found that infants’ persistent food refusal was related to mothers’ concurrent state anxiety, but not to their trait anxiety, which is more general and stable. Consequently, they concluded that maternal anxiety is probably a consequence rather than a cause of child food refusal. However, in the present study, child fussy eating at age 3 and 4 cannot be an antecedent of mothers’ symptoms during pregnancy. Moreover, our results suggest that not only clinically significant anxiety has an effect on child fussy eating, but also slightly elevated anxiety symptoms.
The inclusion of both mothers’ and fathers’ anxiety and depression as contrasting exposures allows us to speculate about underlying mechanisms. Mothers’ anxiety during both pregnancy and the child's preschool period predicted fussy eating in the child. In contrast, fathers’ anxiety during pregnancy was not associated with children's fussy eating. Thus, a genetic explanation is unlikely, whereas these results provide some support for fetal programming.13 The association between fathers’ anxiety during the preschool period and child fussy eating can be explained by parenting factors. Possibly, fathers’ anxiety affects children's fussy eating by controlling feeding practices such as pressure to eat.31 ,32 Such feeding practices could have counterproductive effects by contributing to negative affective reactions to food,33 thereby increasing the risk of food rejection by the child. Parental anxiety may also influence children's fussy eating by affecting difficulties in parent–child interactions.
Like mothers’ depressive symptoms, fathers’ depressive symptoms during pregnancy were related to children's fussy eating. Thus for these associations, fetal programming seems unlikely. Shared heritability of depression and fussy eating could underlie this association pattern, especially bearing in mind genetic influences on fussy eating.34 Possibly, lifestyle or socioeconomic factors affect both parental and child behaviours, although we carefully adjusted for education and income. Parenting factors may mediate the associations of both mothers’ and fathers’ depressive symptoms at 3 years postnatal with child fussy eating. Also, maternal depression has been related to difficulties in the mother–child interaction,35 and, in turn, these problematic interactions could mediate the associations with children's fussy eating.36
Strengths of our study were its large population-based sample, prospective design, and multiple informant ratings. It is noteworthy that our results were similar for maternal and paternal reports of fussy eating at age 3, suggesting that mothers with internalising problems do not over-rate their children's eating behaviour. This also supports the validity of previous findings that relied on mothers’ reports of child eating behaviour.6 ,10 ,11 ,14 However, we did not know whether maternal and paternal reports of fussy eating were completely independent of each other, although two separate questionnaires were mailed. The BSI was used to assess psychiatric symptoms. Although a well-validated instrument,19 its brief character may limit the extent to which it captures all aspects of internalising problems. As with all cohort studies, some selective loss to follow-up among families from low socioeconomic status and non-Western origin occurred in Generation R.16
In conclusion, we observed that maternal and paternal internalising problems were prospectively associated with fussy eating in preschoolers. For effective prevention and management of children's fussy eating, the role of parents’ internalising problems should be considered. Clinicians should be aware that not only severe anxiety and depression, but also milder forms of internalising problems can affect child eating behaviour.
Acknowledgments
The Generation R Study is conducted by the Erasmus Medical Center in close collaboration with the Faculty of Social Sciences of the Erasmus University, the Municipal Health Service Rotterdam area, the Rotterdam Homecare Foundation and the Stichting Trombosedienst & Artsenlaboratorium Rijnmond (STAR). We gratefully acknowledge the contribution of all participating children and their parents, general practitioners, hospitals, midwives, and pharmacies in Rotterdam.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
Press release
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors LMdB: was involved in the conception of the specific research question, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis, wrote the paper, and had primary responsibility for the final content. SCC: provided advice concerning fussy eating behaviour and the interpretation of the data, was involved in drafting the manuscript, and critically revised the manuscript. PWJ: provided consultation for the interpretation of the data, was involved in drafting the manuscript, and critically revised the manuscript. VVWJ: was involved in the design and planning of the Generation R Study and data collection, had overall study oversight, and critically revised the manuscript. FCV: was involved in the design and planning of the Generation R Study and data collection, had overall study oversight, and critically revised the manuscript. OHF: was involved in the design and planning of the Generation R Study and data collection, was involved in the conception of the specific research question, had overall study oversight, and critically revised the manuscript. HT: was involved in the design and planning of the Generation R Study and data collection, was involved in the conception of the specific research question, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis, was involved in drafting the paper, provided consultation for the interpretation of the data, critically revised the manuscript, and had primary responsibility for the final content. AT: was involved in the conception of the specific research question, was involved in drafting the paper, provided consultation for the interpretation of the data, critically revised the manuscript, and had primary responsibility for the final content. All authors read and approved the final manuscript.
Competing interests LMdB, OHF and AT work in ErasmusAGE, a centre for ageing research across the life course funded by Nestlé Nutrition (Nestec Ltd), Metagenics Inc and AXA. The Generation R Study is made possible by financial support from Erasmus Medical Center (EMC), Rotterdam, Erasmus University Rotterdam (EUR), and the Netherlands Organization for Health Research and Development (ZonMw) ‘Geestkracht’ programme (10.000.1003). The work of HT is supported by a NWO-ZonMW VIDI grant (No 017.106.370). The work of PWJ is supported by the Dutch Diabetes Foundation (grant No 2013.81.1664). The funders were not involved in: the study design; collection, analysis and interpretation of the data; writing of the report; and the decision to submit this article for publication.
Patient consent Obtained.
Ethics approval Medical Ethics Committee of the Erasmus Medical Center, Rotterdam.
Provenance and peer review Not commissioned; externally peer reviewed.