The transition from home to child care (e.g., “Kinderkrippe,” “Krabbelstube” or kindergarten)¹ is an exciting, but stressful time in a child’s life generally leading to increased levels of cortisol (Dettling et al., 1999, 2000; Gunnar and Donzella, 2002; Watamura et al., 2003, 2009; Ahnert et al., 2004). At the beginning of “Kinderkrippe” or kindergarten children are faced with a number of new cognitive, behavioral as well as social challenges. They are separated from their parents or other primary caregivers, must comply with the structure imposed by the new child care setting and have to negotiate relationships with a widening circle of peers.

For recently immigrated families the transition to child care can be a particularly challenging experience (Schaich, 2012; Rickmeyer, 2016; Rickmeyer et al., in press). First of all, feelings of loss and separation can be particularly painful and hardly bearable for these families due to their prior migration (Kogan, 2005; Akhtar, 2007; Leuzinger-Bohleber and Fritzemeyer, 2016; Leuzinger-Bohleber and Lebiger-Vogel, 2016). Secondly, an empathic support of the child during transition and associated separation processes might be complicated by potential burdening life circumstances of these families, difficulties in understanding German as well as potential traumatizing experiences and previous losses (Schaich, 2012; Rickmeyer, 2016; Rickmeyer et al., in press).

Cortisol, the primary glucocorticoid of human and non-human primates alike, is the hormonal product of activation of the hypothalamic–pituitary–adrenal (HPA)-axis – a major part of the neuroendocrine system and physiological stress system. It has been well documented that cortisol production increases in response to stress. Experiences with manageable stress are important for a healthy development. However, the beneficial aspects of stress are diminished when it is so severe that it overwhelms a person’s ability to cope with it effectively (Gunnar and Quevedo, 2007). Particularly early in life overwhelming and enduring challenges imposing on the stress system can have negative and prolonged effects on the brain development and parts of the physical stress system (like the HPA-axis). Research increasingly suggests that severe early life stressors may result in decreased brain volumes, elevated brain connections, dysregulation of the neuroendocrine stress response system, and dysfunction of the limbic system involving regions such as the hippocampus, medial prefrontal cortex, and amygdala (Lupien et al., 2009). This in return can compromise proper functioning of the nervous and immune systems and lead to cognitive deficits and emotional dysregulations in adulthood (Gunnar and Donzella, 2002; Tarullo and Gunnar, 2006; Gunnar and Quevedo, 2007; Gunnar et al., 2009; Lupien et al., 2009; Loman and Gunnar, 2010; Matz et al., 2010; Atzil et al., 2011). In accordance to these assumptions, the dysfunction of the HPA-axis is associated with the development of different diseases as well as mental health problems such as depression (e.g., Kirschbaum, 2001; Heim et al., 2008; Stalder and Kirschbaum, 2012).

Prolonged stress early in life can cause a sensitization of the HPA-axis (Gunnar and Quevedo, 2007), leading to an increased reactivity to stressful events and enduring activation of the physiological stress system and thus to a constantly increased release of cortisol (Charmandari et al., 2003; Lupien et al., 2009). However, several studies demonstrated that children who had experienced severe stress during their first years of life, showed elevated cortisol levels during early childhood but decreased cortisol levels later in life (e.g., Susman, 2006; Tarullo and Gunnar, 2006). Therefore several researchers support the so-called attenuation hypothesis. It postulates that in the long run elevations in cortisol due to severe stress early in life lead to a downregulation of the HPA-activity and thus to decreased cortisol levels (Gunnar and Vazquez, 2001; Repetti et al., 2002; Susman, 2006; Trickett et al., 2010).

Normally, the human cortisol follows a circadian rhythm: the cortisol levels are highest at waking, followed by a steady decline across the day, returning to its lowest levels at midnight (Kirschbaum et al., 1999)². Deviations from this pattern are supposed to indicate stress (Kirschbaum and Hellhammer, 1994). A meta-analysis has shown that chronic stress is associated with flatter cortisol slopes across the day, a decreased overall diurnal cortisol level as well as higher cortisol levels in the afternoon and lower levels in the morning (Miller et al., 2007). For example, chronically depressed patients show a decreased or rather flatter decline in cortisol levels across the day (Deuschle et al., 1997; den Hartog et al., 2003). In addition, several studies have demonstrated that lacking positive and supportive relationships with the primary caregivers during childhood can evoke changes of the typical diurnal cortisol pattern, too (Gunnar, 2000; Tarullo and Gunnar, 2006; Roisman et al., 2009). Deviations from the typical diurnal cortisol pattern have been found, for example, in traumatized, neglected and abused children (Deuschle et al., 1997; Gunnar and Vazquez, 2001)³. To conclude, previous research indicates how important it is to reduce stress in early childhood and to support children’s ability to regulate stress effectively. Both is important in order to prevent a permanent activation of the HPA-axis as well as a potentially resulting downregulation of this axis.

Research has demonstrated that the time of child care entry as well as the months afterwards are associated with an increased reactivity of the HPA-axis (e.g., Watamura et al., 2002, 2003; Ahnert et al., 2004; Badanes et al., 2012). Thereby increasing cortisol levels from morning to afternoon, in contrast to decreasing cortisol levels across the day for the same children at home, have been found in several studies (Dettling et al., 2000; Watamura et al., 2003, 2009). The rise of cortisol level across the day at child care seems to be stronger for younger children (36–60 months) than for children of school age (older than 60 months; Dettling et al., 1999; Vermeer and van IJzendoorn, 2006; Bernard et al., 2015). Additionally, children seem to need time to adapt to the new situation. For example, Bernard et al. (2015) demonstrated that children even experienced an increase in the cortisol rise at child care across a 10-week transition phase. Although the elevations in cortisol may decrease after several months spent in child care, even 5 months after starting child care children still show higher levels of cortisol at child care compared to spending the day at home (Ahnert et al., 2004).

Previous work has addressed potential short- and long-term negative effects of full-day care (e.g., Belsky et al., 2007; Roisman et al., 2009; Watamura et al., 2010). For example, Watamura et al. (2010) have demonstrated that children who attended full-day child care and who showed increased salivary cortisol levels had lower antibody levels as well as a higher illness frequency. This work suggests that the rising cortisol pattern may have consequences for physical health (Watamura et al., 2010). Additionally, an analysis of NICHD-data⁴ demonstrated that longer child care hours as well as maternal insensitivity during the first 3 years of life are subsequently associated with lower awakening cortisol levels at the age of 15 (Roisman et al., 2009). The authors concluded that this result is in line with the attenuation hypothesis (Susman, 2006) and may indicate long-term negative effects of both low maternal sensitivity and early full-day child care. However, it still remains unclear what (if any) exact consequences alterations of cortisol level due to child care may have. Furthermore, particularly for children growing up in high-risk environments and experiencing less sensitive parent-child relationships positive early child care experiences may provide a buffer against negative developmental outcomes. Thereby, emotional available teachers could serve as secure alternative attachment figures in terms of a resiliency promoting factor (Werner and Smith, 1982, 1992; Shivers, 2008; Hauser et al., 2009).

Research efforts to understand stress reactivity at child care have demonstrated that variables such as the caregiving quality (Tout et al., 1998; Dettling et al., 2000; Sims et al., 2006; Watamura et al., 2009) as well as the attachment to mothers (Ahnert et al., 2004) and teachers (Badanes et al., 2012) influence the increase of cortisol during child care. It has been postulated that through their contingent and sensitive interactions with their primary caregiver, securely attached children develop regulatory capacities that allow them to modulate stress reactions more effectively. Thus they may be better equipped to deal with the stressors associated with child care entry (Nachmias et al., 1996; Loman and Gunnar, 2010). It is assumed that children who experience enough empathy and a good-enough care and satisfaction of needs through their primary caregivers early in life will develop a basic trust (“Urvertrauen” according to Erikson, 1957) as well as adequate affect regulation (Leuzinger-Bohleber, 2009, 2015). This in return helps them to cope with stressors more effectively later in life (Leuzinger-Bohleber, 2015). In the context of the emotional quality of mother–child relationships the construct of Emotional Availability (EA; Mahler et al., 1975) has often been used to describe the capacity of a dyad to share an emotionally healthy relationship (Emde, 1980; Biringen et al., 2014). It can be assessed with the help of the Emotional Availability Scales (EAS, Biringen, 2008). They measure the affect and behavior of both the child and the mother (caregiver) on six dimensions, four focusing on the caregiver (sensitivity, structuring, non-intrusiveness, and non-hostility) and two focusing on the child (child responsiveness to the caregiver and child involvement of the caregiver; Biringen, 2008; Biringen et al., 2014). The term EA has first been used by Mahler et al. (1975) in order to describe a mother’s supportive attitude and presence in the context of infant/toddler explorations away from her. Furthermore, the descriptions and definitions of EA are based on Bowlby’s (1958) attachment theory which posits, that higher levels of parental EA promote more secure infant attachment and responsiveness to the parent (Easterbrooks and Biringen, 2000; Emde, 2000; Lum and Phares, 2005). Significant findings have confirmed the positive relationship between parental EA and children’s attachment security (see review by Biringen et al., 2014). In addition, various studies have demonstrated that samples with a high risk for developmental problems due to various mental health or psychological issues (e.g., mother’s depression, substance abuse, history of maltreatment, etc.) show comparatively lower EA (Biringen et al., 2014). Additionally, negative associations between parental traumatization and a reduced quality of the parent-child relationship have been found (Schechter et al., 2010; Feeley et al., 2011; van Ee et al., 2012).

In the context of EA and stress responses an interesting study by Kertes et al. (2009) looked at preschoolers’ HPA-axis reactivity (evidenced by rises in cortisol in laboratory to social and non-social threat contexts) in response to a threat. They demonstrated that sensitive parenting helped to regulate stress responses of highly inhibited children to social (but not non-social) threat. Another study suggests that even in the context of a routine caregiving task a higher EA is associated with lower stress reactivity and with earlier circadian patterning in very young infants (Philbrook et al., 2014). The authors investigated both EA and cortisol levels (with the help of saliva samples) in infants at the age of 1 and 3 months. The results demonstrated that infants of more emotionally available mothers showed lower levels of cortisol secretion across the night than infants of less emotional available mothers. Another study, conducted by Little and Carter (2005) in the US, presented an emotional challenge to (mostly) African-American mother-infant-dyads with a low socio-economic status (SES) in order to investigate children’s emotion reactivity. Thereby children of dyads with more emotional available mother–child relationships showed greater emotional control and less difficulties in emotion regulation during the challenging situation.

In the context of kindergarten a study by Kang (2005) provides evidence that EA predicts preschool children’s externalizing and internalizing behavior problems at school. The results suggest that children who had experienced less emotional available mother–child relationships appeared to be less well regulated and adjusted than those with higher emotional available mother–child relationships. Furthermore, in a study by Howes and Hong (2008) examining Mexican-heritage families in the US, maternal structuring as well as sensitivity at home when the child was 3 years of age predicted children’s pretend play, social competence as well as children’s exclusion by peers during pre-kindergarten⁵. Additionally, an investigation showed that multiple dimensions of EA measured in the pre-kindergarten year were associated with kindergarten readiness (Biringen et al., 2005). Thereby, the dimensions mother’s sensitivity and structuring as well as child responsiveness and child involvement, assessed before kindergarten entry, predicted teacher-reported internalizing and externalizing symptoms as well as levels of observed aggression and/or victimization during the transition to kindergarten and at the end of the kindergarten school year.

In order to investigate a potential buffering role of attachment security against elevations of cortisol due to child care entry, Ahnert et al. (2004) have assessed attachment security (with the help of Ainsworth’s Strange Situation paradigm; Ainsworth and Wittig, 1969) as well as cortisol levels of infants and toddlers. The assessments took place during the first and last day of an adaptation phase to child care (parent present at child care), during the first, fifth, and ninth day of separation (parent not present) as well as 5 months after child care entry. The dyads were coded for secure-base behaviors and salivary cortisol was assessed in the morning. The children’s mean age at child care entry was 14.9 months. Only when the mother was present during the adaptation phase securely attached young children showed lower cortisol levels across the day at child care compared to insecurely attached children. However, both securely and insecurely attached children showed increased cortisol levels during the separation phase, when the parent was not present, compared to the adaptation phase. In both attachment groups the cortisol levels were lower after 5 months than during the separation phase, but still higher than home baseline levels. These results indicate that a secure maternal attachment has a buffering role against elevations of cortisol at child care, however, only when the parent is present. This is in line with the assumption that it might be of additional importance in coping with potential stressors whether a responsive caregiver is present or not (Bowlby, 1973; Gunnar and Brodersen, 1992). However, it remained unclear whether the same effects would be found in older children (e.g., at kindergarten age) who may have already developed more stable attachment systems. It could be speculated that older securely attached children are already able to cope with impending stressors more effectively on their own than younger securely attached children.

Badanes et al. (2012) investigated the potential buffering role of both attachment to mothers and teachers in N = 110 children (age 2.0–5.5 years) at child care, with an inclusive sample of both low socio-economic and Mexican-origin families. In their study children with more secure attachments to teachers were more likely to show falling cortisol or less increasing cortisol levels across the child care day. The attachment to mothers did not have an effect on the cortisol pattern. However, maternal attachment interacted with quality of child care, with buffering effects found for securely attached children attending higher quality child care. Accordingly highly secure attached children with high classroom quality showed the optimal decreasing cortisol pattern across the child care day. However, securely attached children with lower classroom quality showed increasing cortisol levels across the day. Furthermore, insecurely attached children showed a flat pattern across the day regardless of classroom quality. Concerning maternal attachment security the authors concluded that low maternal attachment security may prevent the optimal decreasing cortisol pattern even in high-quality child care environments and that securely attached children at this young age are more distressed by lower-quality child care. In summary, this study suggests that after several months spent in child care the attachment to teachers may play a more important buffering role for potential elevations of cortisol in association with child care than maternal attachment (Badanes et al., 2012).

The effect of the emotional quality of the parent-child relationship on children’s elevations in cortisol after child care entry has not been widely studied, yet. The study by Badanes et al. (2012) included both low socio-economic and Mexican families, but was conducted in the US. Ahnert et al. (2004) investigated younger children from middle-class families in Germany. Thus remarkably little is known about cortisol production in response to the beginning of kindergarten among children with an immigrant background in Germany.

However, starting and adjusting successfully to the challenges imposed by kindergarten⁶ can have a positive effect on the child’s development and future school success (Seyda, 2009) and is therefore particularly important for children with an immigrant background who are still disadvantaged in German educational institutions (Autorengruppe Bildungsberichterstattung, 2016). Apart from this, attending kindergarten can have a compensating effect on potential language deficits thus improving the educational opportunities of these children (Becker and Tremel, 2011). However, as mentioned above, due to previous migration-related losses and parents’ potential traumatizing experiences the transition to kindergarten can be a particularly painful experience for families with an immigrant background (Schaich, 2012). In addition, children from these families are more likely to live in high-risk environments (Moro, 2014). In 2014, 44% of children living in families with an immigrant background grew up exposed to at least one situation of risk such as unemployed, low-income earning or educationally disadvantaged parents (Autorengruppe Bildungsberichterstattung, 2016). Thus it is assumed that these children are more likely to have experienced stressful life events than those children who grow up in families with a higher SES and who have not experienced the unique challenges associated with immigration. A review by Mesman et al. (2012) on observational studies of parental sensitivity in ethnic minority families in the US and the Netherlands with young children indicates that parental sensitivity is generally lower in ethnic minority families than in majority families. However, it suggests that the main cause for this difference is family stress due to socio-economic disadvantages. Apart from stress due to socio-economic factors (low SES, unemployment, insecure institutional status, etc.) challenges due to psychological factors associated with the different phases of migration, particularly mothers’ risk of social withdrawal and isolation, can have additional negative effects on the emotional quality of the parent-child relationship (Lebiger-Vogel et al., 2016; Leuzinger-Bohleber and Lebiger-Vogel, 2016). These potential negative effects in turn can have a negative impact on the children’s ability to regulate stressful events. Thus the risk that children with an immigrant background encounter difficulties coping with stressors associated with kindergarten and show increased levels of stress may be significant. Furthermore, potential separation difficulties in the parent-child relationship due to parents’ migration-related losses could additionally complicate the transition to kindergarten (Rickmeyer, 2016; Rickmeyer et al., in press). However, as mentioned before, remarkably little is known about cortisol production in response to the beginning of German kindergarten among children with an immigrant background. Potential associations between their cortisol responses and their relationships with their primary caregivers have not been investigated, yet. Therefore the purpose of the current study was to assess the role of the emotional quality of the mother–child relationship as a potential buffer against cortisol reactivity after kindergarten entry in a sample of children with an immigrant background.

Most of the previous studies demonstrating associations between maternal attachment security and HPA-axis activation have used Ainsworth’s Strange Situation Test (Ainsworth and Wittig, 1969; e.g., Hertsgaard et al., 1995; Ahnert et al., 2004). However, traditional instruments based on attachment theory such as the Strange Situation Test have been criticized for not taking sufficiently into account cultural variations (Rothbaum et al., 2000; Keller, 2012). Therefore in the current study the emotional quality of the mother–child relationship was assessed with the help of the EAS (Biringen, 2008). The EAS have the advantage that they can be rated independently of the caregivers’ cultural background (for an overview see Biringen et al., 2014; Lebiger-Vogel, 2016) with a more dyadic focus than traditional attachment instruments. There is a large body of empirical research using the EAS. They have been applied in more than 100 studies and a lot of different countries all over the world (Biringen et al., 2014).

In the context of parents’ migration and EA, a previous study by van Ee et al. (2012) demonstrated decreased and problematic EA within mother–child relationships of a sample of N = 49 refugees and asylum seekers in the Netherlands. In this study the relationships between maternal posttraumatic stress symptoms, parent-child interaction (using the EAS) as well as infants’ psychosocial functioning and development was investigated. Their results showed a positive relationship between maternal posttraumatic stress symptoms and insensitive, unstructuring or hostile, but not intrusive, parent-child interactions. Furthermore, infants showed lower levels of responsiveness and involvement to their traumatized mothers (van Ee et al., 2012).

In addition, in the current study children’s cortisol levels were assessed with the help of hair cortisol assessment, a new and promising method to measure cortisol (Stalder and Kirschbaum, 2012). Assessments took place approximately 1 week before as well as 3 months after kindergarten entry. HCC assessment offers a measure of cumulative HPA-activity and provides a retrospective reflection of integrated cortisol secretion over periods of several months. In contrast, previous methods to measure cortisol (e.g., saliva, urine, blood collection) are limited in the temporal range of assessment because they provide only snapshots of HPA-activity (Stalder and Kirschbaum, 2012). Only one study published to date in this context has investigated children’s stress level with the help of HCC, yet, however, during the transition to school (Groeneveld et al., 2013)⁷.

The current study was conducted in Frankfurt/Main, Germany, and examined the following questions:

It has been expected that children over the age of three, who have experienced more emotional available maternal relationships, have already developed better regulated stress systems that help them to regulate stress more effectively. Therefore they should show a smaller increase of cumulated HCC after kindergarten entry compared to children with less emotional available mother–child relationships. Thereby the relationship between a potential increase in cortisol and different EA scales was investigated.

All children and families included in the current study had taken part in the early prevention project “FIRST STEPS – An Integration Project for Infants with an Immigrant Background.” The project was conducted by the Sigmund-Freud-Institut (SFI, Frankfurt/Main) in close collaboration with the Anna-Freud-Institut (AFI, Frankfurt/Main) and the IDeA-Center⁸ (for a detailed project description see Lebiger-Vogel et al., 2013, 2015a,b, 2016; Burkhardt-Mußmann, 2015; Rickmeyer et al., 2015; Leuzinger-Bohleber and Lebiger-Vogel, 2016)⁹. It has first been implemented in Frankfurt/Main and then in Berlin and serves as a scientifically evaluated model project. By using a prospective randomized comparison group design the effectiveness of a psychoanalytically oriented intervention (intervention A) is evaluated in comparison to an intervention being offered by lay helpers (intervention B). Both interventions are offered from the time of pregnancy until the commencement of kindergarten. The psychoanalytic approach is in line with a new form of outreaching psychoanalysis (Leuzinger-Bohleber et al., 2011, 2013). This can be viewed as an attempt to transfer psychoanalytic expertise into the field of low-threshold offerings for “children-at-risk” and their families. It has the aim to contribute to early prevention and improving the integration of children deriving from socially marginalized groups of society (Leuzinger-Bohleber, 2014). The mothers who receive intervention A are supported by female psychoanalytically trained project staff¹⁰. The aim of the intervention is to optimize the earliest mother–child relationship by supporting parenting competences as well as mothers’ integration in order to support the emotional, cognitive, and language development of the children. The mothers of the comparison group take part in self-organized mother–child groups led by lay helpers. The project addresses “hard-to-reach” families¹¹, particularly mothers who are either pregnant or recently gave birth (and their husbands) and who have mostly only been living in Germany for a short time and thus have not had the chance to be integrated, yet. In order to evaluate the two interventions a variety of different instruments is applied. Over 300 families have already been recruited in Frankfurt/Main and Berlin. First preliminary EAS baseline data of the FIRST STEPS study (n = 45) indicate that the recruited mother–child dyads show a problematic and decreased EA after birth (assessed 2.5–5.5 months after child birth; see Lebiger-Vogel, 2016).

This study is unique for its thorough observation of parent-child interactions among children with an immigrant background in association with potential elevations in cortisol due to kindergarten entry. Furthermore, to our knowledge elevations in cortisol due to child care entry have not been studied with the help of hair cortisol assessment, yet. Our results indicate, as expected, that children demonstrate higher levels of cumulated cortisol during the challenging time after kindergarten entry compared to the months prior. Thereby two dimensions of EA, mother’s non-intrusiveness and children’s responsiveness, seem to play an important buffering role. Thus taking into account potential negative effects of early stress and associated elevations in cortisol, our results demonstrate how important an empathetic support of children during child care entry (such as kindergarten entry) is. Based on our results it could be assumed that supporting especially both mother’s non-intrusiveness and children’s responsiveness in dyads with an immigrant background could help to decrease potential elevations of cortisol in association with child care entry, thus reducing stress in this critical phase early in life. Furthermore, interventions promoting intercultural parent work and the EA of early child care and kindergarten teachers could facilitate the empathetic support of these children during the transition to child care, too (Shivers, 2008). Thereby trainings that sensitize early child care and kindergarten teachers to the specific challenges of families with an immigrant background in association with child care entry (e.g., due to migration-related losses) could be of particular relevance (Schaich, 2012). Furthermore, research suggests that promoting secure attachment to professional caregivers can be of particular importance for children living in difficult life circumstances and growing up with less sensitive parents in order to prevent negative developmental outcomes (Hauser et al., 2009).

In addition, research among immigrant families is still scarce, but tremendously needed. Not only is there a global increase in immigrants in Germany and other European countries, but children with an immigrant background are still disadvantage in our educational system and are more likely to grow up in high-risk environments (Autorengruppe Bildungsberichterstattung, 2016). The integration of these children is one of the most important challenges for European societies (and other western countries). Supporting their integration is important in order to reduce the risk that these children feel excluded and not wanted by the majority society as well as to prevent or at least reduce social divides, tensions, social withdrawal into subcultural milieus as well as the risk of radicalization (Meurs, 2016).

It is important to note, that on the basis of our results, we cannot conclude that stress due to the transition to kindergarten has either negative or positive effects. The present study only demonstrates that children with an immigrant background showed elevated cumulated cortisol levels in response to kindergarten entry and that this effect seems to be particularly pronounce in less responsive children and children with more intrusive mothers. However, taking into account potential negative effects of cortisol elevations early in life, supporting mother’s non-intrusiveness and children’s responsiveness seems to be a promising approach in order to reduce potential elevations of cortisol in association with kindergarten entry and thus promoting early child development. Previous studies have shown that the EA of a dyad is relatively stable across time, but that it is sensitive to change through interventions supporting emotional available caregiver-child relationships at home as well as in professional caregiving settings (Biringen et al., 2010, 2012, 2014; Baker et al., 2015).

All authors made substantial contributions to conception and design, acquisition of data as well as interpretation of data. CR drafted the article and analyzed the data, while ML-B and JL-V revised the article critically for important intellectual content. Additionally, JL-V supported the ratings of the videos used for data analysis. All authors gave final approval of the version to be submitted and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.