Early maternal perceived stress and children’s BMI: longitudinal impact and influencing factors

Overweight and obesity prevalence, especially among preschool children, has risen dramatically world-wide over the last decades affecting 6.1% of children under five years of age in 2016 [1]. Recent data from the KiGGS study shows that in Germany this fraction is even higher with 9.5% of children age two-six being overweight, of which 2.8% are classified as obese [2]. This is particularly concerning as most of these children will remain overweight in adolescence and adulthood, increasing their risk for co-morbidities like cardiovascular diseases or type 2 diabetes mellitus [3, 4].

Unhealthy diet and physical inactivity have been described as the main risk factors contributing to obesity development [5]. Consequently, child obesity intervention and prevention studies are mainly focusing on implementing changes in eating behavior and physical activity. However, most of these studies failed to reach long-term effects [6, 7], suggesting that other factors such as the living environment and parental behavior play an important role in the context of children’s overweight development. Among others, psychological aspects like early infant parental distress [8] as well as maternal depression [9] emerged as potential factors promoting children’s overweight.

Accumulating research shows the wide-ranging consequences of maternal stress on children’s health, which among others include an increased risk for behavioral problems, asthma, reduced birth weight, and an increased risk for becoming overweight [10–12]. Perceived stress is the individual perception about the stressfulness of life and the ability to handle such stress, which can be influenced by a variety of sources. These so-called stressors include socioeconomic disadvantages as well as recent life events like divorce/separation that can ultimately lead to stress-related physiological dysregulations [13, 14]. In this context maternal stress during pregnancy is known to alter signaling in the hypothalamic-pituitary-axis (HPA) exposing the developing fetus to an excess of glucocorticoids [15], one of the mechanisms discussed to contribute to prenatal growth restriction [16, 17] and an accelerated catch-up-growth increasing the risk for obesity in children’s later life [18–20]. While in the prenatal phase the physiological stress response of the mother can directly affect the child, in the postnatal period - as assessed here in the first two years after birth - the impact of maternal stress on parenting behavior and mother-child interactions become important. There is evidence that changes in feeding styles and practices [21] due to parental or maternal stress can have a significant impact on children’s food composition and energy intake [22]. Especially maternal stress becomes important as mothers often spend significantly more time in direct interaction with the child compared to the fathers. For example higher infant energy intake and increased consumption of breads and cereals during the first six month after birth have been described in association with maternal stress or depression [23], as has children’s reduced consumption of fruits and vegetables [24, 25]. Moreover, stress perceived by children themselves seems to alter their energy intake and food selection with a preference for sweet and high fat foods [25].

The majority of studies conducted so far focused on prenatal or postnatal stress exclusively, while longitudinal maternal stress assessments in relation to children’s weight development are rare. As recently reviewed by Tate et al. [12] and O’Connor et al. [26] most of these studies only used either longitudinal information on maternal stress or children’s overweight development. Therefore, our aim was to investigate the association between maternal stress and body mass index (BMI) trajectories in children in a longitudinal manner, including maternal stress evaluations from pregnancy until children’s age of two years and annual weight assessments of the children up to the age of five years. We complement our study by analyzing which stressors might contribute to the perceived maternal stress level and therefore might have a potential impact also on children’s weight development throughout the years. It has been discussed that the living environment including noise exposure and an unsecure living environment influence weight development in adults [27–29]. Since data on the effects of such stressors on children is sparse [30] we included not only the socioeconomic status of the study participants, but also factors characterizing their living environment such as traffic or residential noise in our analysis.

We hypothesize that with our longitudinal qualitative stress assessment we will be able to identify a time window, in which the child is particularly vulnerable to maternal stress and that such stress exposure experienced in this time window might have a long-lasting effect on the development of overweight in the child.

Obesity is a multifactorial disease with an often protracted onset in childhood and adolescence. Despite the appreciation that high caloric intake and sedentary behavior contribute to overweight development, less attention has been given to the effects of pre- and postnatal perceived maternal stress on weight development in early childhood. Given the few longitudinal studies evaluating this relationship in children [8, 39, 42, 43] our aim was to investigate whether and how maternal stress assessed at different time points in the highly vulnerable early pre- and postnatal period is related to overweight development in preschool children.

Previous studies have suggested that dramatic life events during the prenatal phase, such as the occurrence of a natural disaster or the death of a close relative, can contribute to overweight development in preschool children or in young teenagers [44–46]. In our study we did not observe a longitudinal association of prenatal stress on BMI z-scores but rather identified the perceived stressfulness for the mother during the first year of life as an important factor for an increase in BMI z-scores. This observation is in line with a recent meta-analysis by Tate et al., suggesting that toddlers (1-3 years old) are more vulnerable to maternal stress than infants (<one year) [12]. In addition, as the type of stress evaluated in our study is most likely not comparable in its impact to a natural disaster, the missing association between prenatal perceived stress and BMI z-scores might also at least in part be related to the difference in the type of stress evaluated.

As only maternal stress during the first year after birth had an effect on BMI trajectories we hypothesized that duration of breast feeding, which can be negatively affected by maternal stress, might play a role [47]. However, we did not observe significant differences in breastfeeding duration or the time of solid food introduction in relation to the maternal stress level. This does not exclude the possibility that maternal feeding styles and attitudes may play a role [48, 49] as earlier studies indicate that the parent-created environment can foster obesity-promoting feeding styles and attitudes, which shape the child’s food preferences [50, 51]. In older children however, other social influences and food environments experienced in daycare or preschool may dilute the maternal stress effect. Although this is in line with our observation that only maternal perceived stress during the first year but not thereafter had a persistent effect on BMI z-scores, the design of our study did not allow further evaluation of parental feeding styles and attitudes.

Although high maternal stress levels were associated with higher BMI z-scores in the total LINA cohort, this effect was only present in girls, whereas boys were not affected. Similar observations were made by Suglia et al. [39], who described a higher risk of being obese for five-year-old girls, who had experienced high cumulative stress (including food insecurity, housing insecurity, maternal depressive symptoms, and maternal substance abuse) compared to girls without this experience, a similar effect was missing in boys [39].

In adults and adolescents gender disparity in stress perception and processing has previously been associated to differences in coping mechanisms [52–54]. Children seem to respond in a similar way as in particular girls have been described to respond by impulsive eating, emotional binge eating, and by requesting sweet and high fat foods [25, 39, 55]. In light of these previous findings, our results suggest that already at a very young age changes in eating behavior might play a role in the gender-disparity of BMI-development related to the experience of stress.

In accordance with what we saw for the overall maternal perceived stress, “demands”, “worries” and “lack of joy” had a strong positive association with BMI z-scores in girls only. There are several studies suggesting that maternal depression can promote overweight development in children [56, 57], with indications that this might also be a gender-specific effect [9], e.g. Hernandez et al. reported that maternal depression placed females but not males at a higher risk for obesity at age 18 [58].

On a last scale of our analyses we aimed to characterize potential stressors, which add to the perceived maternal stress level throughout the years. We were able to identify three potential stressors, which were all related to the quality of the living environment (burden due to traffic, residential noise, and poor living conditions) contributing to the overall maternal perceived stress.

We show in this study that different sources of noise including “Residential Noise” - noise from neighbors and pedestrians - and “Traffic” summarizing traffic noise and exhaust, can affect the maternal stress level. Noise exposure is a well-described example of a potentially obesogenic factor, which has already been studied for its impact on prenatal/postnatal growth [59–61] and in association to adiposity and metabolic outcomes in adults [27, 29, 62]. Noise during pregnancy and childhood increased the risk of overweight at age 7, although no association to BMI z-scores was found in this study [63]. Also, an unsecure living environment has been suggested to contribute to overweight development. While Mathis et al. observed that adults who perceive their neighborhood as unsecure were 12 times more likely to be overweight [64], in children neighborhood crime was associated with an increase in weight and a limited outdoor activity [65]. Next to the quality of the living environment the only other stressor with a significant impact on overall perceived maternal stress was a low household income, These stressors are likely related to a low socioeconomic status of the families, which is known to be strongly associated with obesity in the Western world [66, 67].

Interestingly, the stressors themselves did not have a direct effect on girl’s BMI z-scores but rather mediated their effect through their impact on the maternal overall perceived stress level. As the maternal perceived stress is most likely a cumulative account of these stressors and potentially also covers others stress-related factors, single stressors might not have a sufficient predictive power. This is in line with the observation that a combination of adverse effects within a family can increase the obesity risk, whereas unfavorable social factors in isolation often do not [68, 69].

In summary, with our longitudinal qualitative stress assessment we were able to provide evidence for the idea that childhood BMI trajectories develop early and that maternal stress during the first year after birth is a persistent positive predictor of BMI z-scores in girls up to the age of five years.

To reduce the risk for childhood obesity – in particular in girls - behavioral interventions to reduce the mental stress in mothers should be considered in the future.