According to the posttraumatic stress disorder (PTSD) criterion A of the Diagnostic and Statistical Manual of Mental Disorders fifth edition (DSM-5), a traumatic event is a direct or indirect exposure to a life threat or severe injury (1). As parents can perceive a maternal and/or infant life threat during childbirth, this event can fulfil the PTSD criterion A and lead to childbirth-related PTSD (CB-PTSD) (2–4). Up to 33-46% of mothers appraise their childbirth as traumatic (5–8).

CB-PTSD has four symptom clusters: intrusions (e.g., childbirth-related flashbacks and nightmares); avoidance of birth-related cues (e.g., not talking about the birth); negative cognitions and mood (e.g., anhedonia and low mood); and hyperarousal (e.g., increased startle response and hypervigilance towards the infant) (2, 9). Childbirth-related posttraumatic stress disorders has been shown to consist of birth-related and general posttraumatic stress symptoms (10), and researchers have recently argued that CB-PTSD is a distinct sub-type of PTSD (3). Meta-analyses concluded that 3-4% of mothers meet CB-PTSD diagnostic criteria in community samples whereas, in high-risk populations (e.g., emergency caesarean section (ECS)), this increases to 16-19% (11, 12). However, the prevalence of sub-threshold childbirth-related posttraumatic stress symptoms (CB-PTSS) is much higher (12.3% and 21.1%) (13). Evidence showed associations between CB-PTSS and adverse family outcomes (14–16) (17–19).

Given the significant consequences of CB-PTSS on family outcomes (such as breastfeeding, parent-infant bonding, infant development, couple relationship satisfaction) and its economic costs due to increased healthcare utilization (20), early identification of women at risk of CB-PTSS is crucial (12, 21). A screening procedure suitable to the maternity context could facilitate the implementation of early preventive strategies for mothers who otherwise may avoid care in the postpartum period for themselves or their infant. The subjective birth appraisal is a determinant as it is one of the strongest predictors of a traumatic childbirth experience (19, 22–24). An apparently non-traumatic childbirth, as described by healthcare professionals, can be experienced as traumatic by the mother, and vice-versa (19, 22, 25, 26). Therefore, at-risk mothers cannot be identified only based on medical criteria. To date, no validated tool exists to assess maternal risk of developing CB-PTSS shortly after childbirth. Such an acceptable, early screening tool would represent an important clinical opportunity (27).

During childbirth, both lives of mothers and their infants can be threatened, and maternal perceived life threat can contribute to the development of CB-PTSS. While the literature is consistent regarding the predictive nature of maternal perceived life threat perception for oneself (7, 23), the role of a maternal perceived life threat for their infant is less well understood during childbirth. A few studies investigated the role of parental perception of life threat to their child on parental PTSS following other traumatic events (28, 29). For instance, parental perception of the child’s life threat following traffic-related accidents prospectively predicted parental acute stress severity, which was associated with subsequent parental PTSD severity (28). Another longitudinal study with parents of children exposed to a burn observed an impact of parental perceived child’s life threat on intrusion and avoidance symptoms at one-month post-trauma (29). Regarding childbirth, to our knowledge, no study examined the role of maternal perception of infant life threat on maternal CB-PTSS. One experimental cross-sectional study investigated its role on maternal physiological stress reactivity and found that mothers with traumatic childbirth had an altered physiological stress reactivity at three days postpartum, when adjusting for maternal perceived infant life threat during childbirth (30). This supports the assumption of a specific role of the latter in CB-PTSS development. Overall, additional research is needed to better understand the relationship between the perception of threat and CB-PTSS.

This secondary data analysis of two observational longitudinal studies aimed to investigate prospective relationships between the perception of threat, either to one’s own life or to that of one’s infant, and CB-PTSD symptoms at four to six weeks and six months postpartum, in a community sample and an ECS sample. The same associations were explored when applying a cut-off of two or three for the two questions (rated on a Likert scale from 1-7), as the objective was to make the instrument easy to use and interpret for clinicians.

Table 1 shows detailed descriptive information regarding sociodemographic, medical, and mental health characteristics. The life threat and infant threat continuous scores were lower in the community sample than in the ECS sample. Thus, a post-hoc chi-square test was performed to examine the relation between the groups and the response to the life threat (i.e., life threat ≥ 3) and to the infant threat (i.e., infant threat ≥ 3). The relation between these variables was significant, X² (1) = 18.24, p <0.001 and X² (1) = 9.26, p = 0.002 for life threat and infant threat respectively. Thus, women belonging to the ECS sample were more likely to experience a perception of life threat or of infant threat than women belonging to the community sample.

In the community sample, eight percent of mothers perceived a threat to their life during childbirth vs. 40% to their infant’s life. In the ECS sample, 40% of mothers perceived a threat to their life vs. 66.7% to the infant’s. The fact that mothers were up to five times more likely to perceive their infant’s life to be threatened than their own is novel. Some authors have highlighted that the nature of the life threat is often misunderstood, particularly in the context of childbirth. Many studies assessing the traumatic stressor criterion assess both types of threat together, i.e., “Did you think you or your infant’s life was at risk?” (39).

The prevalence of probable self-reported CB-PTSD diagnosis decreased from 5.56% at one month to 0% at six months in the community sample, contrasting with prior findings (4). In the ECS sample, the prevalence rate of probable CB-PTSD increased from 2.67% at six weeks to 6.67% at six months. This is in line with a recent meta-analysis showing an increase between three to four and six months postpartum (13).

In both samples, both types of perceived threat during childbirth were important in predicting maternal CB-PTSS, particularly at four to six weeks postpartum. This is in line with a previous study showing that both types of perceived threat were reported to be higher in women with full CB-PTSD compared to those being non-symptomatic (40).

In the community sample, perceiving any threat during childbirth was prospectively associated with total scores of CB-PTSS and with several CB-PTSD subscales scores. However, in the ECS sample, infant threat was prospectively associated with CB-PTSD outcomes, whereas the life threat yielded no significant association with future CB-PTSS. It is possible that, in this sample, where the objective threat to the infant’s life was increased (most ECS are often a consequence of a threat to the infant’s health (41)), the probability of subjectively perceiving a threat to the infant’s life is also increased (42).

When looking at the CB-PTSD symptom clusters, both types of perceived threat were prospectively associated with hyperarousal, at one or six months, and with avoidance at four weeks in our community sample. In the ECS sample, there were no such similarities: the infant threat score was prospectively associated with most PTSD subscales but not avoidance. One explanation may be that different coping mechanisms are at play in this population, depending on the nature of the perceived threat. If mothers perceived their own life to be at risk during childbirth, avoiding any reminders may be a way of coping. If the infant’s life was perceived to be threatened, mothers may instead cope by being overly alert for signals indicating additional risks for their infant and be influenced by societal expectations and the perception of the maternal role and responsibility to protect infants from harm (43). Indeed, in this ECS sample, a prospective association with hyperarousal at six months was shown for the infant threat but not for the life threat. Infant threat was prospectively associated with intrusion symptoms at both four to six weeks and six months. When using the infant threat, but not the life threat continuous scores. It is thus possible that having feared for the infant during childbirth and subsequently being overly alert to further signals of infant threat may trigger more intrusions symptoms.

The role of covariates is also worth mentioning. In the community sample, the predictive power of life threat decreased when controlling for maternal age and mode of childbirth. This is not surprising, as the mode of childbirth is associated with life threat (44). However, the opposite effect was found when focusing on infant threat whilst controlling for Apgar score and birth weight: more CB-PTSD symptom clusters were prospectively associated with infant threat, both at four to six weeks, as well as at six months.

The strengths of this study include the use of a community as well as an ECS sample and the use of validated questionnaires to measure maternal CB-PTSS and probable CB-PTSD. At the time of data collection, no validated questionnaire to measure CB-PTSS or CB-PTSD existed; therefore, validated generic questionnaires assessing PTSD symptoms were used and their instructions adapted to assess childbirth-related symptoms.

Limitations are the use of different time points to assess life threat and infant threat and the use of different questionnaires to measure CB-PTSS and probable CB-PTSD in mothers across the two samples; thus, conclusions need to be cautiously drawn. More specifically, contrarily to our ECS sample for whom CB-PTSS and probable CB-PTSD were evaluated with the PCL-5, the PDS-F used in our community sample lacks an evaluation of negative changes in mood and cognition, which may be viewed as a limitation in our manuscript. Furthermore, in the covariates, there was missing data, due to the fact that women’s hospital records could not always be accessed. This means that regressions conducted with covariates should be interpreted with caution.

Our results have potential clinical implications: they show that assessing life threat and infant threat in mothers shortly after childbirth via two brief items may help to identify mothers at risk of developing CB-PTSS and probable CB-PTSD up to six months. This study may thus represent the first step in the development of a brief screening tool for CB-PTSS and probable CB-PTSD, which is needed in order to identify those at risk early and to prevent negative consequences for the family (39, 45). Future studies should replicate these results and assess emotional responses to threat, the psychometric properties as well as the acceptability of a brief screening tool.

Not all participants agreed to share their data with persons outside of the START Research Consortium. The ethical study protocol, including the statistical analysis plan, statistical code, and data dictionary are available upon request from the corresponding author.

The studies involving humans were approved by Ethics committee for research in humans of the Canton de Vaud. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

LG: Conceptualization, Data curation, Investigation, Methodology, Writing – original draft, Writing – review & editing, Formal Analysis. VS: Conceptualization, Data curation, Investigation, Methodology, Writing – original draft, Writing – review & editing, Project administration. CD: Conceptualization, Data curation, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing. AH: Conceptualization, Data curation, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing, Funding acquisition, Resources, Supervision, Validation.