Pre-natal Attachment and Parent-To-Infant Attachment: A Systematic Review

During the perinatal period, the establishment of the attachment relationship with the fetus and subsequently with the real child is crucial for the parents' and the child's well-being. Coherently with the assumption that the attachment relationship starts to develop during pregnancy, this systematic review aims to analyze and systematize studies focused on the association between pre-natal attachment and parent-to-infant attachment, in order to clarify the emerging results and provide useful information for clinical purposes. Nineteen studies were included. Sixteen researches identified a positive relationship between pre-natal attachment and parent-to-infant attachment, and three articles highlighted a negative association between antenatal attachment and post-partum bonding disorders. These results were found both in women and men, in normative and at-risk pregnancies, adopting different assessment approaches (i.e., self-report measures, observations, and projective measures). However, only small or moderate associations were found. Future studies are needed to further confirm these findings across different populations (e.g., male samples, non-normative samples or samples in disadvantaged conditions) and with different methodological approaches (e.g., observational measures). Moreover, studies would be needed in order to clarify mechanisms through which pre-natal attachment influences parent-to-infant attachment, as well as protective and risk factors which intervene between these two variables.


INTRODUCTION
The transition to parenthood is a critical period of adjustment for both men and women, where the attainment of the parental role plays a major part in the well-being of the family as well as in the development of the child. During this period, the establishment of an affective tie with the fetus and subsequently the newborn is a major goal that can impact on the child's health and psychological development (Wilson et al., 2000;Laxton-Kane and Slade, 2002;Della Vedova, 2005;Della Vedova et al., 2008;Branjerdporn et al., 2017).
In literature, the term "attachment, " proposed by Bowlby (1969), refers to the relationship that the child develops with a caregiver aiming to obtain a secure base from which to explore the world and, if it is necessary, proximity, protection, and a retreat for safety and comfort in conditions of stress and danger (Ainsworth, 1969;Bowlby, 1969;Ainsworth et al., 1978;Waters and Cummings, 2000). During the first relational experiences, the child develops internal working models (i.e., mental images of the self, others and the relationships) which are the basis of later attachments in life (Bowlby, 1982;Benoit, 2004) and influence affect regulation strategies and cognitions (Brennan and Shaver, 1995;Shaver and Mikulincer, 2002;Mikulincer et al., 2003;Schore and Schore, 2008). According to the "attachment theory" (Bowlby, 1969), at the end of the 1st year of life the child has formed a specific style of attachment to his caregivers. The child's attachment style will be structured as "secure" or "insecure" depending on the quality of early relational experiences. If the baby enjoys a nurturing and safe relationship with those who care for him/her, the baby will gain a secure personal basis and therefore a sense of trust in others. Thus, the complementary parental side of attachment, which Bowlby defined the "caregiving system, " is an indispensable condition for the healthy development of the child (Buist et al., 2003;Righetti and Sette, 2008;Habib and Lancaster, 2010;Della Vedova et al., 2011;Premberg et al., 2011;Velotti et al., 2011;Prino et al., 2016;Vismara et al., 2016;Rollè et al., 2017;Ionio et al., 2019;Riva Crugnola et al., 2020). Ainsworth et al. (1978) identified four dimensions of maternal behaviors which are connected with the security of the infant attachment: sensitivity, acceptance, cooperation and accessibility.
Other researchers (van IJzendoorn, 1995;Miljkovitch et al., 2004;Arnott and Meins, 2007;Korja et al., 2010;Waters et al., 2018) underlined that parental caregiving strategies are influenced by a range of conscious and unconscious factors (i.e., internal working model or attachment's representation) based on first attachment's experiences to their own caregivers. Within this approach, parental attachment's representations of first attachment's experiences influence adult attachment style and affect the sensitivity with which parents respond to their infant's needs (Main et al., 1985;van IJzendoorn, 1992). Therefore, the parental side of attachment is a fundamental condition for the child's development, and, as many authors from a psychoanalytical domain suggested since the '40 (Deutsch, 1945;Winnicott, 1956;Bibring, 1959), it is supposed to start in the mother (and father's) mind during pregnancy. Along with this framework, the interest of clinicians and researchers turned to the pre-natal phases of the relationship, considering the parental affective investment toward the fetus as a precursor of the post-natal relationship. Authors who pioneered this area of research have used the term "attachment, " from Bowlby's theory, to describe the parental emotional bonding with the fetus which gradually develops over the course of pregnancy (pre-natal attachment), extending this concept also to the postnatal period (parent-to-infant attachment). Even though the use of the word attachment to indicate the parental emotional involvement toward the fetus/baby is questionable (Walsh, 2010), pre-natal/post-natal attachment has become commonly used to indicate the affectionate tie that parents develop toward their unborn baby.
Around the '80 a theoretical construct has been created with the aim to systematically investigate the affective investment of the parents toward their fetus. Cranley (1981), on the basis of her experience in the midwifery field, proposed the construct of "prenatal attachment" according to the psychoanalytical literature of maternal adjustment to pregnancy and role attainment (Deutsch, 1945;Benedek, 1959;Bibring, 1959) and created a specific questionnaire to evaluate its characteristics. The Maternal-Fetal Attachment Scale (MFAS; Cranley, 1981), consists of five subscales: differentiation of self from the fetus; interaction with the fetus; attributing characteristics and intentions to the fetus; giving of self; role taking. The measure also includes a parallel form assessing the paternal fetal attachment (PFAS; Weaver and Cranley, 1983). Müller (1993Müller ( , 1996, in order to measure the special affectionate relationship that develops between a woman and her fetus, developed the 21-item Prenatal Attachment Inventory (PAI, 1993) achieved identifying four themes: Preparedness, Fantasizing, Affection, and Interaction. Within this perspective, Condon (1993), starting from the idea that maternal emotional investment toward the fetus must be distinguished from features of maternal pregnancy adaptation and maternal role attainment, reformulated the pre-natal attachment construct. He proposed a model that gives attention to "the emotional bond or tie of affection experienced by the parent toward the infant" (Condon and Corkindale, 1998;p. 57), which from the authors' perspective, promotes dispositions to caregiving that can be subsequently translated into overt behaviors oriented toward the child (i.e., information seeking, proximity, protection, pleasing, and gratifying). The authors defined this tie of affection as parent-tofetus attachment in pregnancy and parent-to-infant attachment in the post-natal period, and identified four indicators of the construct: pleasure in proximity, tolerance, need gratification and protection, knowledge acquisition (Condon and Corkindale, 1998;Scopesi et al., 2004;Condon et al., 2008). To measure the construct, Condon created two questionnaires for mothers and fathers: the Maternal Antenatal Attachment Scale and the Paternal Antenatal Attachment Scale (MAAS, PAAS; Condon, 1993) for the pre-natal period, and the Maternal Post-natal Attachment Scale and Paternal Post-natal Attachment Scale (MPAS, PPAS; Condon and Corkindale, 1998) for the postnatal period.
In the last decades, the terms "parent-infant bonding" or "post-partum bond" were often used as synonyms of "parentto-infant attachment" (Brockington et al., 2006;Altaweli and Roberts, 2010;Bicking Kinsey and Hupcey, 2013). The concept of "bonding" or "bond" has been initially defined by Klaus and Kennell (1976) as a parent-child tie based on skin-to-skin contact during the early sensitive post-partum period. The authors Klaus and Kennell (1982); (Kennell and Klaus, 1998) later criticized their definition as it did not consider the affective domain. According to this, recent literature describes the term "bonding" or "bond" as an affective tie that the parent develops toward the child (Kennell and McGrath, 2005;Taylor et al., 2005;Bicking Kinsey and Hupcey, 2013).
Although, as previously noted, "parent-to-infant attachment" and "parent-infant bonding" has been frequently used as synonyms, nonetheless, regardless of the definition adopted, the affective tie experienced by parents toward the child is the main focus (Bicking Kinsey and Hupcey, 2013).
Therefore, for convenience, we will refer to this specific affective tie in terms of parent-to-infant attachment, independently of the term used in the included studies.
Differences emerged on the assessment approach used to evaluate parent-to-infant attachment and thus on its operationalization (see Table 1), which may influence the obtained results and make it difficult to compare the conclusions highlighted across the studies [for an extensive review see Wittkowski et al. (2020)]. While Condon and Corkindale's MPAS and PPAS (1998), as other self-report instruments -e.g., the MIBS (Taylor et al., 2005) and the PPBS (Cuijlits et al., 2016) -emphasized the affective and cognitive dimension of this construct, focusing on the subjective parental experience of parent-to-infant attachment, other authors assessed it resorting to observational approaches based on behavioral indicators (e.g., Avant, 1982), in line with a large body of literature on infantto-parent attachment, or projective methods (e.g., van Bakel et al., 2013). As stated by Condon and Corkindale (1998), attachment behaviors have to be considered as possible but not necessary consequences of the attachment experience. In line with this consideration and highlighting the methodological biases related to observational approaches, these authors suggest assessing parent-to-infant attachment administering self-report measures, more capable to capture the subjective experience and the core of the attachment bond (i.e., love), which can in turn promote parental dispositions to caregiving. Nonetheless, a combination of observational and self-report measures would be ideally recommended (Condon and Corkindale, 1998;Condon, 2012).
Parent-to-infant attachment emerges as an important factor contributing to the quality of the subsequent reciprocal relationship, as well as to the well-being and the development of the child (Young, 2013;Parfitt et al., 2014). A stronger parentto-infant attachment promotes the child's social, cognitive, and behavioral development (Schenk et al., 2005;Mason et al., 2011;Ip et al., 2018) as well as better mother-child interactions which in turn improve child's emotional regulation (Klaus and Kennell, 1982;Cigoli et al., 2006;Mason et al., 2011;Behrendt et al., 2018Behrendt et al., , 2019Brake et al., 2020;Ponti et al., 2020). In addition, parents with higher levels of parent-to-infant attachment have higher responsiveness and sensitivity, which promote the development of a child's secure attachment and the infant's exploration of the environment maintaining an appropriate level of stimulation (Ainsworth et al., 1971;Solomon and George, 1996;Siddiqui and Hägglöf, 2000;Sandbrook and Adamson-Macedo, 2004;Blair et al., 2006;Rossen et al., 2019).
As stated by several authors (Deutsch, 1945;Winnicott, 1956;Benedek, 1959;Bibring, 1959;Rubin, 1967), the attachment bond starts to develop during pregnancy through the perception of the fetus as a human being with a separate self and specific needs. This relationship between the parent and the fetus begins in the parents' mind on imaginary level through the development of mental representations of the future child (Righetti, 2003;Ammaniti et al., 2006Ammaniti et al., , 2012Raphael-Leff, 2010;Vreeswijk et al., 2014). As pregnancy progresses, this attachment bond includes the imagined child, who is gradually experienced as a separate and differentiate human being through the contribution of fetal movements and fetal ultrasound (Ammaniti et al., 1992;Raphael-Leff, 2001;Della Vedova et al., 2008;Righetti and Sette, 2008;Vreeswijk et al., 2014). Indeed, during pregnancy, the fetus goes through an extraordinary motor and sensory development that allows him/her to perceive and be perceived, and to establish an early psychosomatic communication and first somatic memories (Borsani et al., 2019). Pre-natal attachment has been described as a bond between a parent and his/her unborn child consisting of mental representations, fantasies, emotions and mental capacities necessary to identify another human being (Cranley, 1981;Condon, 1993;Müller, 1993;Laxton-Kane and Slade, 2002;Doan and Zimerman, 2003). In line with the definition of parentto-infant attachment (Condon and Corkindale, 1998), Condon (1993) described this bond as an experience of love expressed through the parental disposition to form an image of the baby, to interact with and protect the fetus while avoiding separation or loss and gratifying his/her needs.
In light of the above-mentioned literature, pre-natal attachment and parent-to-infant attachment appear as noteworthy factors for the quality of the subsequent parent-child reciprocal relationship and for their well-being. Coherently with the assumption that the development of the attachment bond starts during pregnancy, many studies explored the relation between these two variables (see Table 2). In order to clarify the emerging results, the current paper aims to review and systematize the papers focused on the association between pre-natal attachment and parent-to-infant attachment, providing information that can promote early interventions during the perinatal period, which can have long term impacts on relationship quality and familial well-being.

Data Source and Search Strategy
The current systematic review followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement (Moher et al., 2009). Two independent reviewers searched through EBSCO databases (CINAHL Complete, Family Studies Abstracts, Mental Measurements Yearbook, PsycINFO, Social Sciences Abstracts-H.W. Wilson, Sociology Source Ultimate), PubMed, Scopus, and Web of Science (All Databases). They examined titles, abstracts, and full texts to detect eligible studies published until May 2020. The systematic search was performed using the following keywords: ("pre-natal attachment" OR "maternal fetal attachment" OR "parental fetal attachment" OR "paternal fetal attachment" OR "pre-partum attachment" OR "antenatal attachment" OR "pre-natal bond * " OR "maternal fetal bond * " OR "parental fetal bond * " OR Maternal thoughts, feelings, and behaviors toward the child are evaluated Sample items: 1. Quality subscale (9 items, e.g., "Over the last 2 weeks I would describe my feelings for the baby as: dislike (1)intense affection (5)") 2. Absence of hostility toward the infant (5 items, e.g., "When I am caring for the baby, I get feelings of annoyance or irritation: very frequently (1)-never (5)") 3. Pleasure in the interaction with the infant [5 items, e.g., "When I have to leave the baby: I usually feel rather sad (1)-I usually feel rather relieved (5)"] Response options: 1-5 (each item has a 2, 4, or 5 point response option) Post-partum Bonding Questionnaire (PBQ) Brockington et al. (2001) Mother-to-infant attachment Self-report 25 items Four subscales: 1. Impaired bonding 2. Rejection and anger 3. Anxiety about care 4. Risk of abuse Maternal feelings, cognition and behaviors experienced during interactions with the child are evaluated Sample items: 1. Impaired bonding (12 items, e.g., "The baby does not seem to be mine") 2. Rejection and anger (7 items, e.g., "I feel distant from my baby") 3. Anxiety about care (4 items, e.g., "My baby makes me feel anxious") 4. Risk of abuse (2 items, e.g., "I have done harmful things to my baby") Response options: 0-5 Mother-to-Infant Bonding Scale (MIBS) Taylor et al.
Mother-to-infant attachment Mother-to-infant attachment Self-report 16 items No subscales Maternal feelings, cognition and behaviors experienced during interactions with the child are evaluated Sample items: "I feel distant from my baby"; "My baby irritates me"; "My baby is easily comforted" Response options: 0-5

(Continued)
Frontiers in Psychology | www.frontiersin.org Father-to-infant attachment Self-report 19 items Three subscales: 1. Patience and tolerance 2. Pleasure in interaction 3. Affection and pride Paternal thoughts, feelings, and behaviors toward the child are evaluated Sample items: 1. Patience and tolerance (absence of annoyance/irritability, lack of resentment; 8 items, e.g., "When I'm looking after my baby, I feel sad, frustrated or irritated") 2. Pleasure in interaction (satisfaction, competence, involvement, anticipation of the interaction; 7 items, e.g., "When I am with my baby, I feel impatient") 3. Affection and pride (more stable feelings and cognitions as the sense of ownership, pride, and feelings of affection; 4 items, e.g., "In the last 3 months, I felt I have had no time for myself or to do things that I'm interested in") Response options: 1-5 (each item has a 2, 4, or 5 point response option)

Parent-to-infant attachment
Projective method The question presented to the parents is: "Where would you place your newborn baby in your life right now?" Then, with a symbolic sticker, they place the Self of the baby in the sheet's space. The score is represented by the distance between the "Self" circle and the "Self-Baby" circle.
Mother-to-infant attachment Self-report 5 items No subscales Maternal feelings toward the child are evaluated Sample items: "During the last four weeks, I could describe my feeling toward my baby the best as to be loving (item 1), happy and joyful (item 2)" Response options: 0-3 *The construct assessed by each instrument has been defined as "parent/father/mother-to-infant attachment" irrespective of the definition provided by authors.

Inclusion and Exclusion Criteria
Papers were included if they were: (1) an original research paper, (2) published in English, and (3) focused on the association between pre-natal attachment and parent-to-infant attachment. Articles that did not match these inclusion criteria were excluded. Studies that assessed validity, reliability or psychometric properties of pre-natal or post-natal attachment instruments were excluded, considering that the analysis of the association between the two variables was not the main focus of these articles. Reviews and meta-analysis were excluded as well.
No time restrictions for systematic searching were imposed: all the articles published until May 2020 were considered.

Study Selection and Data Extraction
The search on EBSCO provided 1,016 results, 31 of which were selected for the full text review, after the screening of titles and abstracts. The search on PubMed yielded 461 results, 24 of which were selected. Web of Science produced 543 papers and 24 were selected. Scopus provided 2530 articles, 28 of which were selected. After the duplicates' removal, 35 articles were left and were reviewed in their full text. Of these, 19 studies matched the inclusion criteria and were thus included in the current systematic review, while 16 The results showed a positive association between maternal pre-natal attachment and mother-to-infant attachment only during post-partum hospitalization period (not at 1, 4, and 8 months post-partum) among inexperienced mothers, and during post-partum hospitalization, at 1 and 4 months post-partum (not at 8 months) in experienced mothers.

Ferketich and
Mercer (1995) Paternal Paternal fetal attachment was positively associated with father-infant attachment at 6 and 12 months post-natally. Antenatal attachment predicted with 72% and 68% confidence which men will be in the upper and lower attachment quartiles at 6 months and at 12 months post-partum, respectively.

(Continued)
Frontiers in Psychology | www.frontiersin.org Higher levels of antenatal attachment independently predicted higher levels of mother-to-infant attachment.

(Continued)
Frontiers in Psychology | www.frontiersin.org studies were excluded for the following reasons: four articles did not assess the association between pre-natal attachment and parent-to-infant attachment; two were theoretical articles; nine were methodological studies; and one paper was published in a language other than English (Figure 1). Any disagreement between the two independent reviewers during the study selection and data extraction processes were discussed with a third reviewer, and a unanimous agreement was reached.
One study was conducted in Turkey (Daglar and Nur, 2018) and one in Iran (Taffazoli et al., 2015). All the papers were published between 1994 and 2020 (Figure 2). Furthermore, only five studies focusing on the relationship between pre-natal attachment and parent-to-infant attachment were published before 2011 (Mercer and Ferketich, 1994;Ferketich and Mercer, 1995;Müller, 1996;Damato, 2004;Hjelmstedt and Collins, 2008). This lapse in research is noteworthy, since the psychological process in which women become emotionally involved and attached to their fetus in pregnancy had already been theorized and linked to the parent-child relationship after birth during the 1950s by several psychoanalytic theorists (Deutsch, 1945;Winnicott, 1956;Benedek, 1958Benedek, , 1959Bibring, 1959;Bibring et al., 1961).

Methodological Issues
All of the studies included in the current systematic review adopted a quantitative approach based on the administration of self-report instruments (Mercer and Ferketich, 1994;Ferketich and Mercer, 1995;Müller, 1996;Damato, 2004;Hjelmstedt and Collins, 2008;    2020). Only one paper resorted to an observational approach during post-partum (Taffazoli et al., 2015). Each behavior was observed and recorded focusing on three groups of attachment behaviors: emotional behaviors, proximity behaviors and caring behaviors (Taffazoli et al., 2015). Finally, one study assessed both pre-natal attachment and parent-to-infant attachment using a quantitative as well as a projective approach (Fijałkowska and Bielawska-Batorowicz, 2019).
Only Luz et al. (2017) analyzed the influence of pre-natal attachment on one's own parent-to-infant attachment as well as on their partner's through dyadic data analysis.
All of the studies were conducted on parents who were at least 18 years old. Considering socioeconomic characteristics, most participants had a university or a college degree, were employed, had a middle socioeconomic status, were married or involved in a stable relationship, and belonged to an ethnic majority. Only two studies, conducted in Iran and in Turkey, considered unemployed women or housewives and women with a diploma degree (Taffazoli et al., 2015;Daglar and Nur, 2018). The majority of the studies included both primiparous and multiparous parents (Mercer and Ferketich, 1994;Ferketich and Mercer, 1995;Damato, 2004;Dubber et al., 2015;Rossen et al., 2016Rossen et al., , 2017Luz et al., 2017;Daglar and Nur, 2018;Cuijlits et al., 2019;Matthies et al., 2020;Smorti et al., 2020;Zdolska-Wawrzkiewicz et al., 2020).
Lastly, one study analyzed the relationship between the two variables in fathers with an in vitro fertilization (IVF) pregnancy (Hjelmstedt and Collins, 2008) and one paper included a group of women with a twin pregnancy (Damato, 2004).
For a detailed description of the assessment tools used to evaluate parent-to-infant attachment and the dimensions explored see Table 1.
The pre-natal assessment time of the considered variables was homogeneous. In all studies but two, pre-natal attachment tools were administered from the second half of pregnancy, while Rossen et al. (2016Rossen et al. ( , 2017 assessed pre-natal attachment in all the three trimesters of pregnancy. One study did not provide clear information on the assessment time (Zdolska-Wawrzkiewicz et al., 2020). Conversely, the administration time of post-partum tools was heterogeneous. Parent-to-infant attachment was assessed from the postpartum hospitalization period (Mercer and Ferketich, 1994;Ferketich and Mercer, 1995) up to 24 months after childbirth (de Cock et al., 2016).
Regardless of the adopted assessment tools and approaches (i.e., self-reports or behavioral observations), the association between pre-natal attachment and parent-to-infant attachment was confirmed in primiparous and multiparous mothers and fathers, both in advantageous (Ferketich and Mercer, 1995;Müller, 1996;Damato, 2004;Hjelmstedt and Collins, 2008;Condon et al., 2013;Dubber et al., 2015;de Cock et al., 2016;Rossen et al., 2016Rossen et al., , 2017Luz et al., 2017;Petri et al., 2018;Cuijlits et al., 2019;Fijałkowska and Bielawska-Batorowicz, 2019;Matthies et al., 2020;Smorti et al., 2020;Zdolska-Wawrzkiewicz et al., 2020) and disadvantageous socio-economic and sociodemographic conditions (Taffazoli et al., 2015;Daglar and Nur, 2018) as well as in expectant parents in normative and atrisk pregnancies -i.e., IVF and twin pregnancies (Damato, 2004;Hjelmstedt and Collins, 2008). However, while 17 studies identified a significant association between the two variables regardless of the post-natal assessment period (i.e., from the 8th day to the 24th month after childbirth), in the study of Mercer and Ferketich (1994) a significant association was found only considering post-partum hospitalization period (and not at 1, 4, and 8 months post-partum) among inexperienced mothers (those with no previous children) and until the 4th month postpartum (i.e., during post-partum hospitalization, at 1 and 4 months, and not at 8 months) among experienced mothers (those with one or more previous children). Furthermore, Ferketich and Mercer (1995) identified a significant association between prenatal attachment and father-to-infant attachment only until the 4th month post-partum (i.e., during post-partum hospitalization, at 1 and 4 months, and not at 8 months) in experienced fathers and until the 1st month after childbirth (during post-partum hospitalization and at 1 months, and not at 4 and 8 months post-partum) in inexperienced fathers.
In addition to finding a direct association between the two variables among mothers, Smorti et al. (2020) identified a partial mediation through traumatic childbirth experiences. Maternal pre-natal attachment reduced the level of PTSD symptoms associated with traumatic childbirth experiences which in turn promoted a higher quality of mother-to-infant attachment . Furthermore, Damato (2004) showed that the association between maternal antenatal attachment and mother-to-infant attachment in twin pregnancies was moderated by post-partum depression, method of delivery and the need for admission to the Neonatal Intensive Care Unit (NICU). Finally, Luz et al. (2017) found that both paternal and maternal pre-natal attachment were significantly associated with father-to-infant attachment. On the contrary, mother-to-infant attachment was not predicted by paternal pre-natal attachment (Luz et al., 2017).

DISCUSSION
The current systematic review aimed to review and systematize the literature on the association between pre-natal attachment and parent-to-infant attachment, clarifying and discussing the emerging results while reporting relevant information for clinical purposes.
Considering inclusion and exclusion criteria, 19 papers were included. Sixteen studies highlighted a significant and positive association between pre-natal attachment and parentto-infant attachment, and three articles identified a negative association between antenatal attachment and post-partum bonding disorders. The results were confirmed evaluating parentto-infant attachment from post-partum hospitalization to 24 months post-partum, including both women and men, in normative and at-risk pregnancies, irrespective of the used assessment approach (i.e., self-report measures, observations, projective measures). Only Taffazoli et al. (2015) identified a positive relation between pre-natal attachment and emotional and proximity behaviors, while they did not confirm these results with respect to caring behaviors. This inconsistent finding may be a consequence of the cultural impact on parental representation of their caregiving role and of their child's characteristics. Indeed, this is the only study conducted in the Middle East among the reviewed papers otherwise European. Overall, these findings highlight that regardless of the assessment approach used to evaluate parent-to-infant attachment, antenatal attachment is a precursor of attachment bond in the early post-partum, considering both affects, cognitions, behaviors and non-verbal representations as indicators. However, only few studies resorted to observational or projective measures, therefore these results have to be cautiously considered and other studies are necessary.
Nonetheless, these findings support the hypothesis that the attachment bond starts to develop during pregnancy through the development of affects, fantasies and mental representations of the unborn child which influence the relationship with the real child after birth. Pre-natal attachment has an impact on the parents' affects and cognitions as well as on their daily interactions with the child during the early post-partum period, promoting the establishment of early, secure and healthy relationships that were found to be associated with more positive outcomes in the child's development and post-natal infant-toparent attachment (Sroufe, 2005;Murphy and Laible, 2013;Zimmer-Gembeck et al., 2017;Matthies et al., 2020).
However, several studies identified only low or moderate associations between pre-natal attachment and parent-to-infant attachment (e.g., Müller, 1996;Damato, 2004;de Cock et al., 2016;Rossen et al., 2016Rossen et al., , 2017Cuijlits et al., 2019;Matthies et al., 2020). These data may in part depend on the applied assessment tool; not only, other factors may also influence pre-natal and early post-partum attachment, and their link. In line with these findings, some studies (Damato, 2004;Smorti et al., 2020) included in the current systematic review observed that the relationship between antenatal attachment and parent-to-infant attachment was mediated or moderated by gestational childbirth experiences (i.e., cesarean delivery; traumatic childbirth experiences), as well as by the need for neonatal intensive care and psychological symptoms (i.e., postpartum depression). Thus, other perinatal variables seem to intervene in the association between pre-natal and parent-toinfant attachment, and more studies are necessary to clarify these preliminary findings, including individual, relational and contextual variables.
Although the association was widely confirmed among mothers, few studies have been conducted on fathers (Ferketich and Mercer, 1995;Hjelmstedt and Collins, 2008;Condon et al., 2013;de Cock et al., 2016;Luz et al., 2017;Fijałkowska and Bielawska-Batorowicz, 2019). For men, the development and establishment of an attachment relationship with the unborn child presents greater complexities due to the absence of physical evidences and bodily experiences Lancaster, 2005, 2010;Righetti et al., 2005;Ustunsoz et al., 2010;Della Vedova and Burro, 2017). Accordingly, lower levels of pre-natal attachment were found among men compared with women in several studies (Steen et al., 2012;Vreeswijk et al., 2014;Della Vedova and Cristini, 2019;Fijałkowska and Bielawska-Batorowicz, 2019) with paternal pre-natal attachment being influenced by the father's perception of the partner's attitude toward him during pregnancy and his perceived distress (Della . Furthermore, as highlighted in the current systematic review, fathers with a higher pre-natal attachment and whose partners have higher levels of maternal pre-natal attachment show a higher father-to-infant attachment in the first post-natal period (Luz et al., 2017). This suggests that the development and the stability of the attachment relationship in fathers is influenced by individual and relational variables Yeung et al., 2001;Fijałkowska and Bielawska-Batorowicz, 2019), such as the quality and intensity of his partner's attachment to the fetus, highlighting the role of women during the perinatal period for the assumption of the paternal identity (Ferketich and Mercer, 1995;Buist et al., 2003;Condon et al., 2004Condon et al., , 2013Habib and Lancaster, 2010;Luz et al., 2017). However, due to the scarcity of studies on the subject, it is necessary to increase studies on male samples to strengthen and deepen these conclusions.
Furthermore, most of the studies were conducted on specific groups of participants such as parents who were married or involved in a stable relationship, with a college degree, employed, with a middle-class socio-economic status and belonging to an ethnic majority, and parents who experienced normative pregnancies. Therefore, the generalizability of these results is limited.

Clinical Implications
Since all of the studies confirmed the association between prenatal attachment and parent-to-infant attachment, and taking into account that parent-to-infant attachment can impact the well-being of the family system (e.g., the child's development and emotional regulation, parental responsiveness and sensitivity to the child's needs, mother-child interactions), regular screening processes, and preventive programs are suggested at an early stage of pregnancy aiming to increase the strength of the prenatal attachment and to indirectly influence the parent-child relationship after birth (Brandon et al., 2009;Young, 2013;Parfitt et al., 2014;Cataudella et al., 2016). Parental-fetal attachment can be modifiable by specific supporting interventions that emerged as effective (Brisch et al., 2003;Brecht et al., 2012;Akbarzade et al., 2014;Cunen et al., 2017;Ekrami et al., 2019;Parlakian and Kinsner, 2019). These programs -such as psychosomatically oriented antenatal classes, home visiting interventions or prenatal psychoeducation programs -support the development of the attachment bond by providing a psychic space for parenting, promote the parents' awareness of the presence and needs of the child, encourage parents' fantasies about their unborn baby and favor the parent's availability toward their child, perceived as a separate psychological being (Slade, 2005;Ammaniti et al., 2006;Vismara et al., 2020), providing support along the whole perinatal period (Cranley, 1981;Feldman, 2007Feldman, , 2012.

Strengths and Limitations
This review presents several limitations. Firstly, statistical conclusions on the results of this systematic review cannot be drawn as it is not a meta-analysis.
Moreover, this review only included papers published in English, thus excluding results published in other languages that could give a broader understanding of the association between pre-natal attachment and parent-to-infant attachment.
According with the inclusion and exclusion criteria, we only focused on the association between pre-natal attachment and parent-to-infant attachment. However, in future systematic reviews it could be useful to consider the association between pre-natal attachment and infant-to-parent attachment, in order to explore long-term effects of antenatal attachment.
Lastly, we did not consider other factors such as parental mentalization and parental representations which could provide a deeper understanding of the variables that influence relationship quality during the post-partum period.

Future Research Directions
Firstly, further research would be useful to deeply analyze the association between pre-natal attachment and parent-to-infant attachment in male samples and in non-normative pregnancies.
Furthermore, since some studies (Damato, 2004;Smorti et al., 2020) highlighted the role of mediators (i.e., level of PTSD symptoms linked to childbirth) and moderators (i.e., depression, method of delivery and the need for the admission to the NICU) in the association between pre-natal attachment and parent-to-infant attachment, and considering the low or moderate association between the two variables found in many studies (e.g., Müller, 1996;Damato, 2004;de Cock et al., 2016;Rossen et al., 2016Rossen et al., , 2017Cuijlits et al., 2019;Matthies et al., 2020), it could be useful to test more complex models to clarify mechanisms through which pre-and parent-to-infant attachment are connected, and risk and protective factors involved (Sameroff and Fiese, 2000). Future studies investigating the role of parental mentalization in pregnancy as well as the unconscious level of parental representations of caregiving and the child are needed, in order to reach a broader comprehension of psychological and cognitive factors that can influence the relationship quality during early post-partum.
Finally, it is noteworthy that only one study (Taffazoli et al., 2015) used an observational approach, while the remaining papers applied self-report measures, focusing mainly on affective and cognitive indicators of parent-to-infant attachment. Condon and Corkindale (1998) highlighted that behavioral observations of attachment in the early post-partum period can be expensive in research and clinical setting, biased by situational factors and social desirability, and unable to capture the subjective experience of parent-to-infant attachment, recommending using self-report questionnaires able to identify the experiential dimension of attachment. However, combining these two assessment procedures may strengthen the reached conclusions, and may deepen how the affective experience of attachment both during pregnancy and in the post-partum is translated in caregiving behaviors in the early post-natal period. Thus, in future studies it would be important to assess parent-to-infant attachment using both self-report measures and observational procedures (Condon and Corkindale, 1998;Condon, 2012).

CONCLUSION
This systematic review on the association between prenatal attachment and parent-to-infant attachment found a significant association between the two variables. However, few results emerged on male samples, non-normative pregnancies and on disadvantageous socio-economic and demographic samples, suggesting that for these participants results should be interpreted with caution. Future studies are needed to better clarify this relationship and generalize these results.
Nonetheless, these findings can be useful for clinical purposes, providing information for the implementation of screening processes and interventions aimed at enhancing the attachment relationship starting from pregnancy, improving the parent-child relationship and familial well-being during post-partum.

AUTHOR CONTRIBUTIONS
TT and LR took overall responsibility for the creation of the framework used in this review and the selection of the papers. MG, FS, LV, and AMDV searched for the articles discussed in the review. TT and LR supervised the entire work. All authors were involved in the discussion, the writing, and the revision of the manuscript, and they gave the final approval of the version to be published.