Posttraumatic Stress Disorder in Infants and Young Children Exposed to War-Related Trauma

Home , Child development, Maternal sensitivity

NEW RESEARCH

Ruth Feldman, Ph.D., AND Adva Vengrober, M.A.

Objective: Although millions of the world’s children are growing up amidst armed conflict, little research has described the specific symptom manifestations and relational behavior in young children exposed to wartime trauma or assessed factors that chart pathways of risk and resilience. Method: Participants included 232 Israeli children 1.5 to 5 years of age, 148 living near the Gaza Strip and exposed to daily war-related trauma and 84 controls. Children’s symptoms were diagnosed, maternal and child attachment-related behaviors observed during the evocation of traumatic memories, and maternal psychological symptoms and social support were self-reported. Results: PTSD was diagnosed in 37.8% of war-exposed children (n ϭ 56). Children with PTSD exhibited multiple posttraumatic symptoms and substantial developmental regression. Symptoms observed in more than 60% of diagnosed children included nonverbal representation of trauma in play; frequent crying, night waking, and mood shifts; and social withdrawal and object focus. Mothers of children with PTSD reported the highest depression, anxiety, and posttraumatic symptoms and the lowest social support, and displayed the least sensitivity during trauma evocation. Attachment behavior of children in the Exposed-No- PTSD group was characterized by use of secure-base behavior, whereas children with PTSD showed increased behavioral avoidance. Mother’s, but not child’s, degree of trauma exposure and maternal PTSD correlated with child avoidance. Conclusions: Large proportions of young children exposed to repeated wartime trauma exhibit a severe posttraumatic profile that places their future adaptation at significant risk. Although more resilient children actively seek maternal support, avoidance signals high risk. Maternal well-being, sensitive behavior, and support networks serve as resilience factors and should be the focus of interventions for families of war-exposed infants and children. J. Am. Acad. Child Adolesc. Psychiatry, 2011;50(7): 645–658. Key words: PTSD, war-related trauma, toddlers, preschoolers, attachment theory

he understanding that repeated exposure to (PTSD) in nonverbal or minimally verbal chil- war-related trauma can carry a lasting im- dren has similarly posed a range of difficulties. T pact on infants and young children is, sur- Young children often provide incoherent ac- prisingly, a relatively recent phenomenon. Until counts of their experiences, do not have the recently, clinicians and researchers have gener- linguistic capacities to contain and detail over- ally held that young children are resilient to whelming events, and the specific manifestations severe traumatic experiences and lack the cogni- of re-experiencing, dissociation, and hyper- tive mechanisms to both understand the gravity arousal symptoms in young children differ in 1 of such events or hold them in memory. The meaningful ways from those observed in older formal diagnosis of posttraumatic stress disorder children and adults.2,3 As to the specific effects of armed conflict, very little research has systemat- ically addressed the effects of daily exposure to war and terrorism on the development, preva- This article is discussed in an editorial by Dr. Alicia Lieberman on page 640. lence, and symptom formation of PTSD in infants and young children. An interview with the author is available by podcast at www.jaacap.org. Several studies examined PTSD in young children after traumatic experiences, such as Hurri-

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 645 FELDMAN AND VENGROBER

cane Katrina,4 the September 11th terrorist ways to the development of the disorder in some attack,5 natural disasters,6,7 motor vehicle acci- children but not in others. Consistent with the dents,8 or exposure to domestic violence,9-11 tak- view that resilience is a process reflecting natural ing into account the developmental sensitivities, human adaptation,25 war exposure may offer a assessment difficulties, and specific symptom context to describe child and family factors that manifestations at this age group.12 All studies function to enhance positive adaptation in the demonstrated a distinct and detectable PTSD face of chronic stress.26 constellation in young children after traumatic Since the early work of Anna Freud and Dor- experiences that involve a real threat to the othy Burlingham on young children during physical integrity of the self or significant oth- World War II,27 it has been suggested that the ers.13,14 Research on children exposed repeatedly mother’s emotional adaptation serves as a ”de- to war, typically assessing older children or ad- fensive wall” against the effects of war on the olescents, has similarly underscored the long- child and plays an important role in the devel- term effects of wartime experiences on children’s opment of the child’s disorder and in its symp- well-being, emotional reactivity, and mental and tom severity and trajectory over time. Studies physical health.15-17 It has been further empha- assessing children’s traumatic responses to war sized that the assessment of PTSD in young and terrorism across the globe, following the children should be accompanied by direct obser- 9/11 terrorist attack,28 the war in Lebanon,29 vations of the child’s emotional reactions to the military violence in Gaza,30 SCUD missiles evocation of traumatic memories in addition to in Israel,20 the wars in Kosovo31 and Bosnia- maternal reports, in particular, the increase in Hercegovina,17 or the Oklahoma City bombing,32 secure base behaviors or signs of withdrawal and demonstrate that the mother’s posttraumatic numbing, which mark the child’s nonverbal dis- symptoms, depression, and anxiety meaningfully tress to trauma reminders. However, we are increase the risk of childhood PTSD. Similarly, aware of no study that included formal obser- the family’s support networks serve as a buffer vations of young children’s behavior during against maternal psychopathology and increase the evocation of traumatic memories as coded the mother’s ability to contain her own and the by individuals blind to the child’s psychiatric child’s distress. diagnosis. Although little research has focused on ob- Theoretical models of childhood PTSD posit served mother–child interactions, war exposure that the study of risk and resilience after trauma may provide a unique vantage point to study the should be viewed from a perspective that takes attachment system and its ability to function as a into account multiple levels of the child’s ecol- ”secure base”33 during periods of continuous ogy, including the mother’s mental health and stress. Conditions of stress and uncertainty acti- emotional resources, the nature of the mother– vate the attachment system and have tradition- child relationship, and the family’s broader social ally been used to assess individual variability in support networks.2,18-22 In particular, the study its parameters.34 Children experiencing consis- of resilience, defined as positive outcome despite tent maternal sensitivity tend to increase the use significant adversity,23 may benefit from research of secure base behaviors during moments of on the correlates of war-related childhood PTSD. distress, such as proximity seeking or heightened Among the central controversies in the study of focus on mother, which reciprocally elicit more resilience is uncertainty in the measurement of sensitive caregiving.35 In fact, Bowlby’s seminal risk. Not only is it impossible to ascertain the formulations on the attachment system and its accuracy of young children’s reports, but envi- three phases of attachment, separation, and ronments marked by chronic early stress differ loss36-38 were based on observations of young substantially from one another, underscoring the children’s reactions to repeated wartime trauma question of whether resilient children are those and the role of secure base behaviors and the who experienced less adversity.24 War exposure mother’s sensitive response in repairing momen- offers a ”natural experiment” in which all chil- tary disruptions to the attachment relationship. dren are exposed to the same or highly similar Such mechanisms were thought to protect traumatic experiences over a lengthy period; yet, against the avoidance and numbing associated individual differences in the child’s biological with the loss of attachment bonds, which results and social provisions may chart specific path- from the mother’s prolonged inability to function

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 646 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

as a secure base. In support, animal and human and daily living domains, and would exhibit studies have demonstrated that maternal prox- more behavioral avoidance during the evocation imity and sensitive caregiving provide ”external of traumatic memories, whereas Exposed-No- regulators” that buffer against the hyperarousal, PTSD children would display more secure base anxiety, and bio-behavioral disorganization behaviors during trauma evocation. (3) Regard- caused by maternal separation, whereas pro- ing maternal risk and resilience factors, we ex- longed separation and the decompensation of the pected that mothers of PTSD children would mother’s regulatory functions results in hypoac- report higher depression, anxiety, and posttrau- tivity and biological and behavioral withdraw- matic symptoms and lower social support as al.39-41 It can thus be expected that sensitive compared with mothers of Exposed-No-PTSD mothering and children’s increased use of secure children, whereas mothers of Exposed-No-PTSD base behaviors would chart a resilience pathway children would display greater sensitivity during and may buffer against the development of trauma evocation. Finally, (4) attachment-related PTSD. On the other hand, the avoidant behav- maternal and child behavior during the evoca- ioral response typical of loss would mark a risk tion of traumatic memories—maternal sensitiv- pathway and correlate with greater posttrau- ity, child secure base behavior, and child matic symptomatology in young children. avoidant behavior—would each be indepen- Avoidance as a signal of more pathological reac- dently predicted by components of the child’s tion to war-related trauma has similarly been ecology, including maternal, child, and contex- reported in older children30 and adults.42 tual factors. In light of the above, the current study examined the development of PTSD in infants and young children 1.5 to 5 years of age exposed to METHOD war-related trauma over a lengthy period. A Participants large cohort of Israeli children living at the bor- Participants included 232 children 1.5 to 5 years of age der of the Gaza Strip and exposed to daily (mean age ϭ 33.08 months, SD ϭ 10.89 months) and ϭ ϭ rockets and terrorist attacks were observed with their mothers (mean age 31.27 years, SD 5.55 ϭ their mothers and compared with nonexposed years, range 22.3–47.4 years). Of the children, 47.6% were male and 47.1% were firstborn. Maternal and matched controls. Two major goals guided the paternal education averaged 3.88 (SD ϭ 1.44) and 3.51 study: (1) to describe the PTSD constellation in (SD ϭ 1.49) respectively on a scale of 1 (elementary infants and young children following war- school) to 4 (college education). Among mothers, related trauma in terms of specific symptoms, 31.6% worked full-time and 42.6% worked part-time, developmental regression, and observed attach- and 11% were single mothers. ment-related behavior, including secure base and A total of 148 families comprised the war-exposed avoidant behavior during the evocation of trau- group. These included children living in the same matic memories, and (2) to assess maternal cor- neighborhoods in the town of Sderot, located 10 km relates of childhood PTSD related to both ob- from the Gaza border. Citizens of Sderot were exposed served maternal sensitivity and mental health repeatedly to rocket attacks over a period of several years, had only 15 seconds to enter protected spaces factors. We sought to chart maternal and child after hearing the alert sirens, and were exposed to risk and resilience pathways to the development frequent mortar shelling to which no alert signals were of PTSD and thus, a special focus were the provided. Testing was conducted during a period of differences between war-exposed children who repeated rocket and missile attacks (January 2006– developed PTSD and those exposed to the same October 2008). During this period rocket attacks on wartime trauma who did not develop the disor- Sderot occurred unpredictably and continuously sev- der (Exposed-No-PTSD children). eral times a month, and all children were seen at least Our specific hypotheses were: (1) a subgroup 1 month after the period these attacks began. Visits of war-exposed children would develop the early were not conducted during the day of the attack or within the next 2 to 3 days. childhood PTSD syndrome and exhibit the typi- Recruitment was conducted through clinicians liv- cal symptoms in the re-experiencing, avoidance, ing in Sderot or in neighboring cities who were famil- and hyperarousal domains. (2) With regard to iar with the clinical and childcare services for this age; child risk and resilience factors, we hypothesized once recruitment had begun, participants helped iden- that children with PTSD would show greater tify eligible friends and neighbors. Advertisements for developmental regression in the cognitive, social, study participation were posted in all childcare loca-

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 647 FELDMAN AND VENGROBER

tions across Sderot, and the great majority of families Home visits began with the clinician interviewing approached (Ͼ90%) agreed to participate. It can thus the mother and collecting detailed information for the be acssumed that the majority of young children diagnosis of PTSD, including the specifics of the growing up in these frontline neighborhoods partici- trauma exposure, the child’s and family members’ pated in the study and exclusion criteria included only degree of exposure, and the child’s emotional reaction those who were physically or mentally handicapped to specific events, including description of incidences (e.g., severe autism, mental retardation). Family re- of exposure, the child’s proximity to the explosion, ceived $80 in vouchers for their participation. whether the child or family members were injured and A total of 84 nonexposed children were recruited as how severely, and the child’s behavioral expression of controls from towns within the greater Tel-Aviv area fear, horror, and other emotions in response to each (e.g., Or Yehuda, Pardes Katz) that were equivalent to traumatic exposure (criterion A). The clinicians then the town of Sderot in terms of population size, socio- elicited the mother’s detailed description of the child’s economic composition, and housing and employment specific posttraumatic symptoms. The re-experiencing opportunities. These areas were not exposed to war- domain (criterion B) is typically observed in young related trauma during the study period, and controls children through the expression of trauma reminders were matched to the exposed group in age, gender, in words or gestures during play or daily activities, birth order (firstborn/later born), maternal and pater- compulsive or repetitive play that re-creates elements nal age and education, and maternal employment and of the traumatic events, or repeated thoughts, flash- marital status. Before home visits, control families backs, or freezing in response to trauma reminders. were screened by phone for major traumatic events in The avoidance domain (criterion C) is often observed the child’s life (e.g., terror exposure, motor vehicle at this age through a set of behaviors indicating accidents, physical or sexual abuse), and those report- disengagement, such as social withdrawal, constric- ing such trauma were excluded. The study was ap- tion of affective range, lack of interest in daily activi- proved by the University’s Institutional Review Board, ties, or avoidance of people or places that remind the and all mothers signed informed consent. child of the trauma. Based on previous research at this age,3,8 only one symptom in the avoidance category is Procedure required to diagnose PTSD in young children. The Child PTSD. Trained clinicians with a background in hyperarousal category refers to behavioral and physi- early childhood development and psychopathology ological agitation expressed in difficulties in falling visited families at home and diagnosed the child’s asleep or remaining asleep, concentration problems, PTSD using the Diagnostic Classification: Zero-to- exaggerated startle, hypervigilance, tantrums or unex- Three Revised (DC:0-3R).43 Clinicians received exten- plained anger outbursts, or quick mood shifts; two sive training on the clinical features of PTSD at this age symptoms are required in this category for the diag- and its specific behavioral manifestations; how to nosis of PTSD at this age. Information on a fourth interview mothers and elicit information on young category of symptoms, fears and aggression, was col- children’s emotional state; the specific symptom man- lected, as suggested by the DC 0-3R criteria but is not ifestations at this age; how to approach mothers and required for the diagnosis of PTSD. Similarly, as per children living under continuous war-related trauma; the guidelines, detailed information was collected on the specific DC 0-3R criteria for diagnosing early the child’s developmental regression, defined as a childhood PTSD and the information that should be skill the child had mastered but lost after an intense period collected on the nature and proximity to the traumatic of trauma exposure, and mothers were asked about devel- event, child emotional reaction, specific symptoms in opmental regression in the various domains. the various symptom clusters; and how to provide a After the clinical interview, mothers were asked to developmentally sensitive evaluation of the child’s rate the child’s symptoms on a list of 58 posttraumatic developmental regression. Clinicians were supervised symptoms that were compiled after a year of pilot by a senior clinical child psychologist and a child study in Israeli and Palestinian war-exposed locations. psychiatrist, and cases were conferred every few The pilot included extensive interviews conducted weeks. with parents, caregivers, and therapists living in war The home visit was intended to diagnose the child zones on the typical symptoms young children exhibit in his/her natural ecology and to enable the observa- after repeated exposure to war-related violence.44 The tion of maternal and child behavior during the evoca- final list of 58 symptoms included symptoms detailed tion of traumatic memories. Such observations have by the DC 0-3R as typical of young children after been advocated as central for the diagnosis of PTSD in trauma and additional symptoms described by parents young children.2, 3 Information was collected during and clinicians during the pilot. For each symptom, one afternoon visit that lasted approximately 3.5 to 4 each mother rated if she had not noticed the symptom hours (with breaks). A second visit was scheduled to in her child (score ϭ 0) and if she did, whether the collect questionnaires and additional information if child exhibited the symptom infrequently (1 ϭ once a data were missing. week or less) or frequently (2 ϭ daily or every 2–3

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 648 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

days). Scores were then summarized into the ”re- nal style, appropriate range of affect, reflective ability experiencing” (7 items), ”avoidance” (13 items), ”hy- on the child’s state and behavior, coherence of narra- perarousal” (11 items), ”fears and aggression” (13 tive, awareness of the child’s distress, and maternal items), and ”developmental regression” (14 items). supportive presence. Child secure base behavior includ- Possible score ranges were as follows: re-experiencing, ed: regressive behavior (e.g., thumb-sucking), child 0 to 14; avoidance, 0 to 26, hyperarousal, 0 to 22; and reliance on mother (e.g., ”Get me this..”), child focus fears and aggression, 0 to 26. on mother and increased vigilance, proximity-seeking In terms of developmental regression, mothers first behaviors, whine/cry/call to mother, and child cud- rated the child’s overall regression on a scale of 1 to 10 dling on mother (␣ ϭ 0.84). Child avoidant behavior and then rated specific symptoms of developmental included child gaze aversion, emotional withdrawal, regressions in the social (e.g., staying alone at a moving away from mother’s arms’ reach, and in- friend’s house, four items), emotional (e.g., intense fear creased and inappropriate preoccupation with objects when put to sleep, five items), and daily-living (e.g., (␣ ϭ 0.81). toilet training, five items) domains, which were summed for each category. Control mothers were interviewed in the same manner. Mothers were asked Statistical Analysis to describe events the child experienced as ”traumatic” Multivariate analysis of variance (MANOVA) with or emotionally difficult during the past few months group (PTSD, Exposed-No-PTSD, and control) and and both interview and symptom list followed the child gender as the between-subject factors examined same procedure as the exposed group. Mothers de- differences in three clusters, i.e., child PTSD symp- scribed mildly ”traumatic” events in the child’s life, toms, maternal psychological symptoms, and maternal such as loss of a grandparent, illness in the family, or and child interactive behavior. Post hoc comparisons moving to a new neighborhood. with Scheffé tests followed significant main effects. Maternal Psychopathology. Mothers’ depression was Three hierarchical regression equations predicted ma- assessed with the Beck Depression Inventory,45 mater- ternal and child attachment-related behavior (maternal nal anxiety with the State-Trait Anxiety,46 and mater- sensitivity, child secure base behavior, and child nal PTSD symptoms with the Post-traumatic Diagnos- avoidant behavior) from maternal, child, and contex- tic Scale.47 Mothers also completed the Social Support tual factors, and Pearson’s correlations assessed asso- 48 ciations between the predictor variables. The sample Scale. 54 Maternal and Child’s Behavior During the Evocation provides sufficient power to detect small effect sizes. of Traumatic Memories. During the maternal PTSD interview, which lasted approximately 1 hour, children were present in the room and a trained assistant RESULTS followed the child and videotaped the mother and PTSD Syndrome in Young Children child’s behavior throughout the interview. Tapes were Of the 148 children exposed to repeated war-related later coded for maternal and child behavior during trauma, 56 were diagnosed with PTSD (37.8%) trauma evocation. Maternal behavior, emotional state, and 92 were classified as Exposed-No-PTSD. No narrative coherence, and awareness of the child’s dis- child in the control group was diagnosed with tress, and the child’s emotions and behaviors to PTSD, ␹2 ϭ 262.4, p Ͻ.000. No differences in trauma reminders were coded offline by raters blinded to maternal and child psychiatric information along 35 demographic factors emerged between children scales (20 for mother and 15 for child) each rated from who did or did not develop PTSD apart from ϭ 1 (low) to 5 (high). The instrument was based on the child age: PTSD children were older (mean Coding Interactive Behavior,49 a well-validated coding 36.22 months, SD ϭ 10.36 months) than Exposed- system for adult and child behavior during interac- No-PTSD children (mean ϭ 31.50 months, SD ϭ tions. The system has shown good psychometric prop- 10.18 months), F (df ϭ 1, 147) ϭ 6.97, p ϭ .009. We erties and sensitivity to adult and child interactive thus measured differences in the prevalence of behavior related to age, culture, biological and social- PTSD in toddlers versus preschoolers. Among emotional risk conditions, and the effects of interven- 50-53 the 92 exposed toddlers (1.5–3 years), 24 were tion. The maternal sensitivity construct of the CIB diagnosed with PTSD (26.08%). However, among has been associated with the child’s attachment behav- the 56 exposed preschoolers (3–5 years), 25 re- ior, including secure base and avoidant behavior, 2 41 ceived a diagnosis of PTSD (44.6%), ␹ ϭ 5.81, during a separation–reunion episode. Interrater reli- ϭ ability was conducted for 46 (20%) interactions and p .016. These findings indicate a significantly reliability averaged 93%, intraclass r ϭ 0.91. greater risk for developing PTSD in exposed Three composites were created on the basis of preschoolers as compared with toddlers. previous research. Maternal sensitivity (␣ ϭ 0.82) in- Children presented a range of posttraumatic cluded the following items: consistency of the mater- symptoms unique to this age group. MANOVA

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 649 FELDMAN AND VENGROBER

TABLE 1 Posttraumatic Symptoms, Developmental Regression, Maternal Symptoms, and Observed Behavior in War- Exposed Children With Posttraumatic Stress Disorder (PTSD), Exposed Children Without Posttraumatic Stress Disorder (Exposed-No-PTSD), and Controls

Exposed-No-PTSD PTSD (a) (b) Controls (c)

M SD M SD M SD Univariate F

Posttraumatic symptoms Re-experiencing 7.52 3.20 4.62 1.34 1.73 0.80 22.04***aϾbϾc Avoidance 15.61 3.47 8.27 4.56 1.86 1.10 34.16***aϾbϾc Hyperarousal 17.89 5.73 11.83 4.66 6.36 3.22 23.52***aϾbϾc Developmental regression Overall regression 7.78 2.92 3.55 1.62 1.32 0.90 18.54** aϾbϾc Emotional 4.87 3.14 2.66 1.87 1.24 1.03 10.39***aϾbϾc Social 4.94 2.88 2.59 2.20 1.78 1.44 11.76***aϾbϾc Daily living 3.65 2.60 2.29 2.03 1.73 1.30 12.68***aϾbϾc Maternal psychological symptoms Depressiona 8.10 7.81 6.31 5.84 3.71 2.91 10.54***aϾbϾc Anxietyb 40.18 10.73 37.68 9.56 33.79 7.92 8.79***aϾbϾc Posttraumatic symptomsc 6.82 11.80 3.64 7.47 0.58 3.01 11.36***aϾbϾc Maternal and child behavior during trauma evocation Mother sensitivityd 3.02 0.64 3.28 0.55 3.53 0.56 7.19** cϾbϾa Child secure base behaviord 2.54 0.83 2.93 0.98 2.23 0.86 5.23* bϾaϾc Child avoidant behaviora 2.18 0.75 1.54 0.79 1.13 0.52 10.62***aϾbϾc

Note: Information on child PTSD symptoms and developmental regression were based on a maternal questionnaire of 58 items including symptoms described by the Diagnostic Classiﬁcation of Mental Health and Developmental Disorders of Infancy and Early Childhood (DC 0-3R) and symptoms compiled in war-exposed Israeli and Palestinian locations. M ϭ mean. aMeasured with the Beck Depression Inventory.45 bMeasured with the State-Trait Anxiety Inventory.46 cMeasured with the Posttraumatic Diagnostic Scale. dCoded on a scale of 1 (low) to 5 (high) using the Coding Interactive Behavior Manual.49 *p Ͻ .05, **p Ͻ .01, ***p Ͻ .001.

assessing children’s symptoms in the three social life; and had difﬁculties in sleep, mood symptom clusters (re-experiencing, avoidance, regulation, and self-quieting. hyperarousal) with group (PTSD, Exposed-No- Although all children experienced the same PTSD, control) and child gender as the between- traumatic events, the degree of direct exposure subject factors showed an overall main effect for differed (e.g., some homes were hit while mothers group, Wilks F (df ϭ 6, 441) ϭ 7.86, p Ͻ.000, effect or child were present). Child proximity to the size (ES) ϭ 0.11. Univariate analyses with post traumatic event correlated with the number of hoc comparisons show that PTSD children pre- child avoidance symptoms, r ϭ 0.29, p Ͻ. 01, but sented the most symptoms in all symptom clus- not with other types of symptoms. Similarly, the ters as compared with Exposed-No-PTSD chil- mother’s proximity to the event correlated with the dren, who showed signiﬁcantly more symptoms child’s avoidance symptoms, r ϭ 0.31, p Ͻ. 01. Chil- in all clusters than controls (Table 1). No child dren in the PTSD and Exposed-No-PTSD groups did gender effects were found. not differ in the degree of proximity to the trauma. Figure 1 presents the most prevalent symp- However, mothers of PTSD children reported greater toms in each symptom cluster that were observed proximity to the traumatic event (mean ϭ 3.87, SD ϭ in more than 60% of children diagnosed with 1.24) than mothers of Exposed-No-PTSD children PTSD (23% of exposed children). As seen, most (mean ϭ 2.78, SD ϭ 2.21), F (df ϭ1,147) ϭ 4.37, p Ͻ .05. PTSD children re-created the trauma in play, No correlations were found between the number of speech, and gestures; showed social withdrawal, child symptoms or PTSD diagnosis and the number increased interest in objects, and constriction of of days since the last attack.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 650 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

FIGURE 1 Percentages of children diagnosed with war-related posttraumatic stress disorder (PTSD) exhibiting prevalent symptoms in the re-experiencing, avoidance, and hyperarousal categories.

Finally, because of differences in the defini- post hoc comparisons indicated that PTSD chil- tion of preschool PTSD between the DC:0-3R dren showed greater developmental regression and the alternative criteria,11 according to in all domains compared with Exposed-No-PTSD which there is no need for expressing intense children, who showed significantly more regres- fear and horror during the traumatic event sion than controls who displayed little or no (A2), the re-experiencing expressed in repeti- regression (Table 1). Analysis of variance tive play is not necessarily trauma related (ANOVA) assessing the child’s overall regression (criterion B), and only one hyperarousal symp- revealed that PTSD children showed significantly tom is required for diagnosis (criterion D), we greater regression than Exposed-No-PTSD chil- examined the prevalence of PTSD according to dren, who showed greater regression than con- the alternative criteria. This resulted in the trols (Table 1). inclusion of 14 additional children in the PTSD MANOVA assessing the three attachment- group, leading to a total of 70 children diag- related behavior scores (i.e., maternal sensitivity, nosed with PTSD (47% of exposed group). child secure base behavior, and child avoidant behavior) showed an overall main effect for Child Risk and Resilience Factors group, Wilks F (df ϭ 6, 441) ϭ 3.98, p Ͻ.001, ES ϭ MANOVA assessing the child’s developmental 0.08. With regard to child factors, univariate tests regression in the three domains (emotional, so- (Table 1, Figure 2) with post hoc comparisons cial, daily living) with group and child gender as showed that Exposed-No-PTSD children showed the between-subject factors showed an overall the highest levels of secure base behavior, control main effect for group, Wilks F (df ϭ 6, 441) ϭ children scored lower, and PTSD children exhib- 4.27, p Ͻ.001, ES ϭ 0.08. Univariate analysis with ited the least amount of secure-base behaviors.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 651 FELDMAN AND VENGROBER

FIGURE 2 Child behavior during the evocation of traumatic memories. Note: Coded on a scale of 1 to 5 using the Coding Interactive Behavior Manual49 from observations during the evocation of traumatic memories. PTSD ϭ posttraumatic stress disorder. *p Ͻ.05, **p Ͻ.01.

However, PTSD children displayed the highest symptom in each symptom-cluster. Maternal level of avoidant behavior, Exposed-No-PTSD PTSD correlated with the child’s overall post- children showed less, and controls exhibited little traumatic symptoms, r ϭ 0.25, p Ͻ.01, and with or no avoidant behavior. the number of the child’s avoidance symptoms, r ϭ 0.33, p Ͻ.001. In terms of observed maternal behavior, moth- Maternal Risk and Resilience Factors. ers of control children showed the highest levels MANOVA assessing maternal psychological of sensitivity, mothers of Exposed-No-PTSD chil- symptoms (depression, anxiety, PTSD) with dren showed lower, and mothers of PTSD chil- group and child gender as the between-subject dren exhibited significantly lower sensitivity factors showed an overall main effect for group, than mothers of Exposed-No-PTSD children Wilks F (df ϭ 6, 441) ϭ 5.22, p Ͻ.000, ES ϭ 0.07. (Table 1). Univarite tests (Table 1) with post hoc comparisons showed that mothers of PTSD children scored the highest on symptoms of depression, Predicting Maternal and Child Attachment-Related anxiety, and PTSD, mothers of Exposed-No- Behaviors PTSD children scored lower, and mothers of Three regression equations were computed pre- controls scored the lowest (Figure 3). ANOVA dicting maternal sensitivity, child secure base with post hoc comparisons indicated that moth- behaviors, and child avoidance behavior. Predic- ers of PTSD children had lower social support tors included maternal, child, and contextual (mean ϭ 17.96, SD ϭ 10.79) than mothers of factors for a full ecological model, were the same Exposed-No-PTSD children (mean ϭ 19.84, SD ϭ in each regression, and were entered in five 10.99), whose social support did not differ from blocks. In the first block, demographic factors controls (mean ϭ 22.42, SD ϭ 9.62), F (df ϭ 1, 231) ϭ were entered, including child age (in months), 3.83, p ϭ .023. Finally, we examined correlations maternal age, and maternal education, to partial between maternal PTSD symptoms with the out variance related to demographic conditions. child’s overall symptoms and the number of In the second block, the mother’s mental health

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 652 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

FIGURE 3 Maternal depression, anxiety, and posttraumatic stress disorder (PTSD) symptoms in mothers of war- exposed children with PTSD, war-exposed children without PTSD, and controls. Note: Numbers represent z scores. Maternal anxiety was assessed with the State-Trait Anxiety Inventory,46 maternal depression with the Beck Depression Inventory,45 and maternal PTSD with the Posttraumatic Diagnostic Scale.47 *p Ͻ.05, **p Ͻ.01.

factors were entered: depression, anxiety, and diagnosis predicted additional variance above PTSD. In the third block, the family social sup- and beyond all factors in the model. port was entered. In the fourth block, the child Before computing the regressions, correlations overall developmental regression was entered. In were examined between the predictor variables the ﬁfth block, the child’s PTSD diagnosis was and signiﬁcant correlations are reported. Mater- entered to examine whether child psychiatric nal education correlated with greater social sup-

TABLE 2 Predicting Maternal Sensitive Containment and Child Avoidance Behavior During the Evocation of Trauma

Criterion Maternal Sensitivity Child Secure Base Behavior Child Avoidant Behavior

Predictors Change Beta R2 Change F Change Beta R2 Change F Change Beta R2 Change F Change

Child age Ϫ0.06 Ϫ0.24§ 0.22§ Mother age 0.09 0.07 Ϫ0.05 Mother education Ϫ0.24§ 0.06 5.37** 0.10 0.05 4.18§ 0.03 0.04 3.86§ Mother depression Ϫ0.18§ Ϫ0.22§ 0.27** Mother anxiety Ϫ0.27 Ϫ0.18 Ϫ0.22 Mother posttraumatic Ϫ0.33§ 0.10 7.42** Ϫ0.26§ 0.08 5.58§ 0.36** 0.11 6.93** symptoms Social support 0.19§ 0.03 3.82§ 0.30** 0.06 4.82§ Ϫ0.12 0.02 2.85 Child developmental Ϫ0.25§ 0.04 4.31§ Ϫ0.22 0.03 3.88§ 0.23 0.03 3.98§ regression Child PTSD Ϫ0.16 0.01 2.43 Ϫ0.32** 0.06 5.83§ 0.29§ 0.03 4.02§

2 ϭ Ͻ ϭ Ͻ ϭ Ͻ ϭ Note: R total 0.25, F 9, 218 7.86, p .001. 0.28, F9, 216 8.36, p .001 0.23, F9, 216 7.11, p .001. PTSD posttraumatic stress disorder. §p Ͻ.04; **p Ͻ.01.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 653 FELDMAN AND VENGROBER

port (r ϭ 0.27, p Ͻ.001) and less maternal depres- traumatic memories. Child diagnosis was exam- sion (r ϭϪ0.21, p Ͻ.01) anxiety (r ϭϪ0.32, ined from an ecological perspective and in- p Ͻ.001), and posttraumatic symptoms (r ϭ cluded, in addition to maternal reports, struc- Ϫ0.19, p Ͻ.05). Maternal age correlated with less tured observations of behavior. Theoretically, anxiety (r ϭϪ0.22, p Ͻ.01). Maternal depression war exposure was viewed as a ”natural experi- correlated with anxiety (r ϭ 0.62, p Ͻ.001) and ment” that enabled a unique context to examine maternal PTSD symptoms (r ϭ 0.32, p Ͻ.001), risk and resilience pathways differentiating chil- and maternal anxiety and PTSD were interre- dren who developed the full-blown disorder lated (r ϭ 0.25, p Ͻ.001). Social support was from those exposed to the same trauma who associated with lower depression (r ϭϪ0.40, were more resilient. The present results may thus p Ͻ.001) and maternal PTSD symptoms (r ϭ contribute to the general discussion on resilience Ϫ0.26, p Ͻ.001). Finally, child developmental and shed light on the functioning of the attach- regression correlated with greater maternal ment system under conditions of chronic stress PTSD symptoms (r ϭ 0.24, p Ͻ.01). The three and uncertainty. behavior scores were interrelated: Maternal sen- It is of interest that some of the major devel- sitivity correlated with more secure base behav- opmental theories that shaped clinical thought ϭ Ͻ ϭ ior (r 0.27, p .001) and less avoidance (r were based on the observations of young chil- Ϫ Ͻ 0.26, p .001), and children’s secure base be- dren exposed to war-related trauma during havior correlated with less avoidant behavior World War II. These include Anna Freud’s con- ϭϪ Ͻ (r 0.33, p .001). ceptualizations on ego development and de- Results of the three regressions (Table 2) indi- fenses,27 Bowbly’s seminal formulations on at- cate that all models were significant and the tachment security and disorganization,33 and predictor variables tested here cumulatively ex- Spitz’s descriptions of early regressive states and plained 25%, 28%, and 23% of the variance in children’s anaclitic depression and nonorganic mother sensitivity, child secure base behavior, failure to thrive.55 Combined with the early ani- and child avoidant behavior respectively. Inde- mal research of Hofer on maternal proximity as pendent predictors of maternal sensitivity in- providing a set of bio-behavioral hidden regula- cluded maternal education, less maternal depres- tors39 and current neurobiological perspectives sion and PTSD, greater social support, and less on childhood trauma,56 it is clear that continuous child developmental regression. Independent war not only exerts a profound effect on chil- predictors of child secure base behavior included dren’s biology and behavior but that such setting lower child age, less maternal depression and may provide a unique context to study resilience PTSD, greater social support, and no PTSD diag- as a process reflecting natural human adapta- nosis. Independent predictors of child avoidant tion.25 Unlike conditions of early adversity such behavior included higher child age, maternal as poverty or maltreatment, the cohort of chil- depression, maternal PTSD, and child PTSD di- dren living in frontline neighborhoods and ex- agnosis. posed to the same or similar wartime stressors represents a variety of family adaptation, from excellent to very poor, and a range of attachment DISCUSSION histories. As such, differentiating the symptoms Although the numbers of young children around that spell risk from those that mark a more the world growing up amidst armed conflict resilient course and charting the maternal, child, appear to increase each decade, very little is contextual, and attachment-related correlates of known about the psychological growth, posttrau- resilience may be of theoretical and clinical im- matic symptoms, and observed behavioral man- portance. ifestations of infants and young children exposed Overall, the data indicate that a large propor- to war-related trauma over a lengthy period. This tion of young children (37.8%) growing up in war study is among the first to provide a detailed zones and exposed to daily shooting, rockets, or account of the early childhood PTSD constella- missile attacks are likely to develop PTSD, with tion in a large cohort of children exposed to war frequent symptoms of inconsolability and agita- and terror in terms of specific symptoms, devel- tion, repeated expression of trauma reminders in opmental regression, and attachment-related ma- daily life, and substantial constriction of social ternal and child behavior during the evocation of life. Such children also tend to present a marked

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 654 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

developmental regression in the social, emo- ers of exposed children without PTSD, indicating tional, and daily living domains. Specific post- that the mother’s resilience in the face of trauma traumatic symptoms observed in more than 60% shapes the child’s psychiatric condition. Moreover, of the diagnosed children included nonverbal results indicate that the degree of maternal prox- re-experiencing of trauma in words and gestures; imity to the traumatic event, not the child’s prox- frequent crying, night waking, and mood shifts; imity, differentiated war-exposed children with and symptoms of social withdrawal, such as and without PTSD, highlighting the critical impact preference for solitary functional play and in- of the mother’s posttraumatic state on the child’s creased interest in objects combined with de- disorder, as has long been suggested by Anna creased interest in people. These findings may Freud.27 The cross-generation associations between suggest that more than a third of the world’s maternal and child’s posttraumatic symptoms fur- young children exposed to armed conflict over a ther point to the close links between maternal lengthy period are likely to present a severe posttraumatic psychopathology and the develop- enough psychiatric profile that places their future ment of childhood PTSD. These findings are in adaptation at significant risk and requires intense accordance with much previous research in older intervention. Between the ages of 1.5 and 5 years, children and adolescents on the effects of maternal substantial maturation occurs in environment- psychopathology on the child’s reaction to war, dependent brain systems implicated in executive natural disasters, or severe accidents.17,59,60 The control and stress modulation; children enter the effect of maternal symptoms on the child’s prone- social world; and marked strides are noted in the ness to psychopathology is likely transmitted domains of language, symbol formation, empa- through both genetic vulnerabilities and concrete thy, and behavior regulation.57-58 The disruptions maternal behavior. As seen, mothers of Exposed- to the maturation of these skills caused by expo- No-PTSD children were able to provide more sen- sure to repeated war trauma place the future sitive containment to the child’s distress during adjustment of such infants at great risk. trauma evocation, expressed in consistent and pre- Several factors differentiated war-exposed dictable style, appropriate range of affect, reflective children who did or did not develop PTSD: child capacity, awareness of the child’s distress, and the age, maternal psychopathology, family social provision of supportive and calming presence to support, and maternal and child attachment- the child’s fears and anxieties. Maternal sensitivity related behaviors. With regard to child age, ex- at moments of distress is considered the corner- posed preschoolers were nearly twice as likely to stone of attachment theory36 and is thought to develop PTSD compared with exposed toddlers. provide a regulatory framework for the develop- Although these findings require much further ment of bio-behavioral stress-management sys- research, it is possible that the preschool years tems.35 Consistent with much research in the at- mark an especially vulnerable period for young tachment tradition,35 less maternal depression and children exposed to chronic stress or trauma. At posttraumatic symptomatology each accounted for this stage, children gain linguistic, symbolic, and unique variance in maternal sensitivity, underscor- executive skills that enable them to project to the ing the links between maternal emotional state and future, thereby markedly increasing the child’s sensitive behavior. In addition to mental health, anxiety, and improve their ability to express fears mothers of exposed children without PTSD re- in words, play, and actions.57-58 Similarly, by the ported greater social support and their support preschool years the child has already mastered networks did not differ from those reported by important milestones and the developmental re- controls. Models on childhood PTSD18 consider gression may thus be more notable. On the other social support as an important ecological asset hand, as compared with school-age children, that increases the family’s resilience during peri- preschoolers have not yet developed the formal ods of extended trauma. At the same time, it is operative thought, the more refined self-regula- also possible that more resilient mothers are tory repertoire, or the focus on social life and the better able to recruit a larger system of support to ability to draw comfort from the peer group, provide a holding environment for themselves placing preschool children 3 to 5 years of age at and their children. especially high risk. Of interest were the differences in children’s Mothers of PTSD children reported more symp- attachment behavior during trauma evocation, toms of depression, anxiety, and PTSD than moth- which may chart attachment-based mechanisms

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 655 FELDMAN AND VENGROBER

for the development of PTSD in young children. tive approach behavior for the processing of Although the more resilient Exposed-No-PTSD trauma. Second, the cross-generation associa- children increased their secure base behavior dur- tions between maternal PTSD and the child’s ing the evocation of traumatic memories, PTSD avoidance symptoms point to mechanisms of children exhibited high levels of avoidant behavior. modeling and social learning. Mothers of PTSD According to Bowbly,36-38 during moments of dis- children, who themselves avoided the active re- tress children with a history of sensitive caregiving working of the trauma, modeled the use of increase their secure base behavior, including prox- avoidant defenses to the child’s traumatic expe- imity seeking, heightened vigilance, crying for riences.21 Similar ﬁndings are reported for Pales- mother’s attention, and regressive behavior. These tinian children exposed to repeated military acts high-arousal approach behaviors aim to elicit in Gaza, which indicated that children whose greater maternal sensitivity, which, in turn, helps to mothers suffered posttraumatic symptoms and re-establish the child’s physiological and behav- experienced the highest levels of exposure to ioral equilibrium. It thus appears that the child’s military aggression presented the highest levels high arousal and regressive behavior mark a resil- of avoidance symptoms.61 These data highlight ient, rather than risk pathway, which may seem child avoidance—expressed in both daily symp- counter-intuitive to mothers and caregivers. Inter- toms and concrete behavior in response to trauma vention programs for war-exposed families should reminders—as an indictor of high risk and are thus assist mothers in perceiving the child’s regres- consistent with the adult PTSD literature, which sive needs as a sign of strength and coping and indicates that symptoms of avoidance and numb- learn to respond to the child’s increased arousal ing are markers of a more severe manifestation of with more sensitive containment. the disorder and chart a worse prognosis.62,63 Child avoidance, on the other hand, emerged Maternal and child attachment-related behav- as the most notable risk pathway in trauma- iors were independently predicted by compo- exposed young children. During the evocation of nents of the child’s ecology, including maternal, traumatic memories PTSD children increased child, and contextual factors, thus supporting their avoidant behavior, including gaze aversion, ecological models of childhood PTSD.18 Mothers emotional withdrawal, increased preoccupation who were less educated, more traumatized and with objects, and physical distancing from the depressed, had lower social support, and had mother, and such children also exhibited the children who displayed greater developmental highest level of avoidance symptoms. Consistent regression were less sensitive during trauma with Bowbly’s formulation on loss,38 avoidance evocation, and such diminished sensitivity may may indicate that the attachment system has not have further exacerbated the child’s symptoms. been able to contain the child’s anxieties, that Younger children whose mothers were less de- initial attempts to recruit the maternal supportive pressed, had fewer posttraumatic symptoms, and presence have failed, and that the child has shut received more social support were able to ex- down biologically and behaviorally. The link press more secure base behavior. Finally, between a more severe course of the mother’s avoidant behavior was more prevalent among posttraumatic condition and the child’s avoid- preschoolers as compared with toddlers, was ance is repeatedly seen in the data. Maternal predicted by maternal depression and PTSD, and proximity to the traumatic event and the moth- was related to the child’s PTSD diagnosis. These er’s PTSD symptoms correlated with the child’s ﬁndings highlight the mother’s depression, post- avoidance symptoms, not with other types of traumatic symptoms, and social support net- symptoms, and maternal depression and PTSD works as central to the formation of her own and predicted the child’s avoidant behavior. Two the child’s attachment-related behavior during potential mechanisms may thus be suggested on extended periods of stress and emphasize the the links between maternal posttrauma and child need to provide interventions to address the avoidance. First, mothers who were more trau- emotional needs of mothers raising young chil- matized and depressed exhibited less sensitive dren in areas exposed to war-related trauma. containment of the child’s distress, anxiety, and It is important to note that because of the lack regressive needs and these children may have of longitudinal data, it is not possible to conclude learned to internalize and ignore their anger, that maternal psychopathology had a causal ef- fear, or dependency rather than use more adap- fect on the development of the child’s PTSD.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 656 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011 WAR-RELATED PTSD IN EARLY CHILDHOOD

Similarly, because mother and child’s attachment the face of daily trauma, and may assess the behaviors were assessed after a lengthy exposure interplay between the parenting envelope and to trauma and during the evocation of traumatic the child’s biology in shaping the distinct symp- memories, it is not possible to determine the tom constellation defined as posttraumatic stress child’s attachment classification before trauma disorder of infancy and early childhood. & onset. These limitations should thus be consid- Accepted March 2, 2011 ered in the interpretation of the findings. Future Dr. Feldman and Ms. Vegrober are with Bar-Ilan University, Ramat- intervention and longitudinal studies are needed Gan, Israel. to provide further insights on the effects of This study was supported by a National Alliance for Research on trauma on the functioning of the attachment Schizophrenia and Depression independent investigator award (R.F.). system. Disclosure: Dr. Feldman and Ms. Vengrober report no biomedical financial interests or potential conflicts of interest. Further research is required to detail the ef- Correspondence to Dr. Ruth Feldman, Department of Psychology fects of continuous war, terror, and armed con- and the Gonda Brain Sciences Center, Bar-Ilan University, Ramat- flict on young children’s mental health, behavior Gan, Israel 52900; e-mail: [email protected] regulation, and developmental trajectories. Fu- 0890-8567/$36.00/©2011 American Academy of Child and Adolescent Psychiatry ture research may explore genetic, hormonal, and DOI: 10.1016/j.jaac.2011.03.001 temperamental risk and resilience pathways in

REFERENCES 1. Benedek EP. Children and psychic trauma: a brief review of 15. Allwood MA, Bell-Dolan D, Husain SA. Children’s trauma and contemporary thinking. In: Eth S, Pynoos R, eds. Posttraumatic adjustment reactions to violent and nonviolent war experiences. J Stress Disorder in Children. Washington, DC: American Psychi- Am Acad Child Adolesc Psychiatry. 2002;41:450-457. atric Press; 1985:1-16. 16. Elbedour S, Onwuegbuzie AJ, Ghannam J, Whitcome JA, Abu 2. Chu AT, Lieberman AF. Clinical implications of traumatic stress Hein F. Posttraumatic stress disorder, depression, and anxiety from birth to age five. Annu Rev Clin Psychol. 2010;6:69-494. among Gaza Strip adolescents in the wake of the second Uprising 3. Scheeringa MS, Zeanah CH, Cohen JA. PTSD in children and (Intifada). Child Abuse Negl. 2007;31:719-729. adolescents: toward an empirically based algorithma. Depress 17. Smith P, Perrin S, Yule W, Rabe-Hesketh S. War exposure and Anxiety. In press. maternal reactions in the psychological adjustment of children 4. Scheeringa MS, Zeanah CH. Reconsideration of harm’s way: from Bosnia-Hercegovina. J Child Psychol Psychiatry. 2001;42: onsets and comorbidity patterns of disorders in preschool chil- 395-404. dren and their caregivers following Hurricane Katrina. J Clin 18. Pynoos R, Steinberg A, Wraith R. A developmental model of Child Adolesc Psychol. 2008;37:508-518. childhood traumatic stress. In: Cicchetti D, Cohen D, eds. 5. Coates S, Schechter D. Preschoolers’ traumatic stress post-9/11: Developmental Psychopathology. Vol 2. NewYork: Wiley; relational and developmental perspectives. Psychiatr Clin North 1995:72-95. Am. 2004;27:473-489. 19. Koenen KC, Moffitt TE, Poulton R, Martin J, Caspi A. Early 6. McFarlane AC. Posttraumatic phenomena in a longitudinal study childhood factors associated with the development of posttrau- of children following a natural disaster. J Am Acad Child Adolesc matic stress disorder: results from a longitudinal birth cohort. Psychiatry. 1987;26:764-769. Psychol Med. 2007;37:181-192. 7. Ohmi H, Kojima S, Awai Y, et al. Posttraumatic stress disorder in 20. Laor N, Wolmer L, Mayes LC, et al. Israeli preschoolers under pre-school aged children after a gas explosion. Eur J Pediatr. Scud missile attacks. A developmental perspective on risk-mod- 2002;161:643-648. ifying factors. Arch Gen Psychiatry. 1996;53:416-423. 8. Meiser-Stedman R, Smith P, Glucksman E, Yule W, Dalgleish T. 21. Pfefferbaum B. Posttraumatic stress disorder in children: a review The posttraumatic stress disorder diagnosis in preschool- and of the past 10 years. J Am Acad Child Adolesc Psychiatry. elementary school-age children exposed to motor vehicle acci- 1997;36:1503-1511. dents. Am J Psychiatry. 2008;165:1326-1337. 22. Scheeringa MS, Gaensbauer TJ. Posttraumatic stress disorder. In: 9. Cohen JA, Mannarino AP. A treatment outcome study for sexu- Zeanah CH, eds. Handbook of Infant Mental Health. 2nd Ed. ally abused preschool children: initial findings. J Am Acad Child New York: Guilford Press; 2000:369-381. Adolesc Psychiatry. 1996;35:42-50. 23. Rutter M. Parental mental disorder as a psychiatric risk factor. In: 10. Lieberman AF, Van Horn P, Ozer EJ. Preschooler witnesses of Hales R, Frances A, eds. American Psychiatric Association An- marital violence: predictors and mediators of child behavior nual Review. Vol 6. Washington, DC: American Psychiatric Press; problems. Dev Psychopathol. 2005;17:385-396. 1987:647-663. 11. Scheeringa MS, Zeanah CH, Myers L, Putnam FW. New findings 24. Luthar SS, Cicchetti D, Becker B. The construct of resilience: a on alternative criteria for PTSD in preschool children. J Am Acad critical evaluation and guidelines for future work. Child Dev. Child Adolesc Psychiatry. 2003;42:561-570. 2000;71:543-562. 12. Chemtob CM, Nomura Y, Abramovitz RA. Impact of conjoined 25. Masten AS. Ordinary magic. Resilience processes in develop- exposure to the World Trade Center attacks and to other trau- ment. Am Psychol. 2001;56:227-238. matic events on the behavioral problems of preschool children. 26. Luthar SS, Sawyer JA, Brown PJ. Conceptual issues in studies of Arch Pediatr Adolesc Med. 2008;162:126-133. resilience: past, present, and future research. Ann N Y Acad Sci. 13. Graham-Bermann SA, Howell K, Habarth J, Krishnan S, Loree A, 2006;1094:105-115. Bermann EA. Toward assessing traumatic events and stress 27. Freud A, Burlingham D. Children in War. New York: Medical symptoms in preschool children from low-income families. Am J War Books; 1943. Orthopsychiatry. 2008;78:220-228. 28. Chemtob CM, Nomura Y, Rajendran K, Yehuda R, Schwartz D, 14. Scheeringa MS, Zeanah CH, Drell MJ, Larrieu JA. Two ap- Abramovitz R. Impact of maternal posttraumatic stress disor- proaches to the diagnosis of posttraumatic stress disorder in der and depression following exposure to the September 11 infancy and early childhood. J Am Acad Child Adolesc Psychia- attacks on preschool children’s behavior. Child Dev. 2010;81: try. 1995;34:191-200. 1129-1141.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 50 NUMBER 7 JULY 2011 www.jaacap.org 657 FELDMAN AND VENGROBER

29. Bryce JW, Walker N, Ghorayeb F, Kanj M. Life experiences, 47. Foa EB. Posttraumatic Stress Diagnostic Scale Manual. National response styles and mental health among mothers and children in Computer Systems Inc; 1995. Beirut, Lebanon. Soc Sci Med. 1989;28:685-695. 48. Crockenberg S. Predictors and correlates of anger toward and 30. Qouta S, Punamanaki R, Sarraj E. Mother–child expression of punitive control of toddlers by adolescent mothers. Child Dev. psychological distress in war trauma. Clin Child Psychol Psychi- 1987;58:964-975. atry. 2005;10:135-156. 49. Feldman R. Coding interactive behavior (CIB) manual. unpub- 31. Lopes Cardozo B, Kaiser R, Gotway CA, Agani F. Mental health, lished manuscript. Bar-Ilan University; 1998. social functioning, and feelings of hatred and revenge of Kosovar 50. Feldman R, Masalha S. Parent-child and triadic antecedents of Albanians one year after the war in Kosovo. J Trauma Stress. children’s social competence: cultural specificity, shared process. 2003;16:351-360. Dev Psychol. 2010;46:455-467. 32. Pfefferbaum B, Nixon SJ, Tucker PM, et al. Posttraumatic stress 51. Feldman R, Eidelman AI, Sirota L, Weller A. Comparison of responses in bereaved children after the Oklahoma City bombing. skin-to-skin (kangaroo) and traditional care: parenting outcomes J Am Acad Child Adolesc Psychiatry. 1999;38:1372-1379. and preterm infant development. Pediatrics. 2002;110:16-26. 33. Bowlby J. A Secure Base; Parent-Child Attachment and Healthy 52. Feldman R, Keren M, Gross-Rozval O, Tyano S. Mother-Child Human Development. New York: Basic Books; 1988. touch patterns in infant feeding disorders: relation to maternal, 34. Bell SM, Ainsworth MD. Infant crying and maternal responsive- child, and environmental factors. J Am Acad Child Adolesc ness. Child Dev. 1972;43:1171-1190. Psychiatry. 2004;43:1089-1097. 35. Cassidy J, Shaver PR. Handbook of Attachment: Theory, Re- 53. Feldman R, Granat A, Pariente C, Kanety H, Kuint J, Gilboa- search, and Clinical Applications. New York: Guilford Press; Schechtman E. Maternal depression and anxiety across the post- 1999. partum year and infant social engagement, fear regulation, and 36. Bowlby J. Attachment and Loss. Vol 1: Attachment. New York: stress reactivity. J Am Acad Child Adolesc Psychiatry. 2009;48: Basic Books; 1969. 919-927. 37. Bowlby J. Attachment and Loss. Vol 2: Separation. New York: 54. Cohen J. A power primer. Psychol Bull. 1992;112:155-159. Basic Books; 1973. 55. Spitz R. Anaclitic depression. Psychoan Study Child. 1946;2:313- 38. Bowlby J. Attachment and Loss. Vol 3: Loss. New York: Basic 342. Books; 1980. 56. Teicher MH, Andersen SL, Polcari A, Anderson CM, Navalta CP, 39. Hofer M. Hidden regulators: implication for a new understanding Kim DM. The neurobiological consequences of early stress and of attachment, separation, and loss. In: Goldberg S, Muir R, Kerr childhood maltreatment. Neurosci Biobehav Rev. 2003;27:33-44. J, eds .Attachment Theory: Social, Developmental, and Clinical 57. Blair C, Diamond A. Biological processes in prevention and Perspectives. Hillsdale, NJ: Analytic Press; 1995:203-230. intervention: the promotion of self-regulation as a means of 40. Feldman R. Maternal versus child risk and the development of preventing school failure. Dev Psychopathol. 2008;20:899-911. parent-child and family relationships in five high-risk popula- 58. Feldman R, Eidelman AI. Biological and environmental initial tions. Dev Psychopathol. 2007;19:293-312. conditions shape the trajectories of cognitive and social-emotional 41. Feldman R. Maternal-infant contact and child development: in- development across the first years of life. Dev Sci. 2009;12:194- sights from the kangaroo intervention. In: L’Abate L, Embry D, 200. Baggett S, eds. Low-Cost Approaches to Promote Physical and 59. Ostrowski SA, Christopher NC, Delahanty DL. Brief report: the Mental Health: Theory, Research, and Practice. New York: impact of maternal posttraumatic stress disorder symptoms and Springer; 2007:323-351. child gender on risk for persistent posttraumatic stress disorder 42. Engdahl RM, Elhai JD, Richardson JD, Frueh BC. Comparing symptoms in child trauma victims. J Pediatr Psychol. 2007;32:338- posttraumatic stress disorder’s symptom structure between de- 342. ployed and nondeployed veterans. Psychol Assess. 2010; in press. 60. Spell AW, Kelley ML, Wang J, et al. The moderating effects of 43. Zero To Three. Diagnostic Classification of Mental Health and Devel- maternal psychopathology on children’s adjustment post-Hurri- opmental Disorders of Infancy and Early Childhood: Revised Edition cane Katrina. J Clin Child Adolesc Psychol. 2008;37:553-563. (DC:0-3R). Washington, DC: Zero to Three Press; 2005. 61. Qouta S, Punamäki R, El Sarraj E. Prevalence and determinants of 44. Feldman R, Vengrober A, Hallaq E. Mother-child relationship, PTSD among Palestinian children exposed to military violence. child symptoms, and maternal well-being in israeli and palestin- Eur Child Adolesc Psychiatry. 2003;12:265-272. ian infants and young children exposed to war, terror, and 62. Beck KD, Jiao X, Pang KCH, Servatius RJ. Vulnerability factors in violence. Paper presented at the 13th meeting of the European anxiety determined through differences in active-avoidance be- Society for Child and Adolescent Psychiatry, Florence, Italy, 2007. havior. Prog Neuropsychopharmacol Biol Psychiatry. 2010;34: 45. Beck AT. Beck Depression Inventory. San Antonio, TX: Psycho- 852-860. logical Corporation/Harcourt Brace Jovanovich; 1978. 63. Foa EB, Stein DJ, McFarlane AC. Symptomatology and psycho- 46. Spielberger C, Gorsuch R, Lushene R. State-Trait Anxiety Scale. pathology of mental health problems after disaster. J Clin Psychi- Palo Alto, CA: Consulting Psychologists Press; 1970. atry. 2006;67(Suppl 2):15-25.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY 658 www.jaacap.org VOLUME 50 NUMBER 7 JULY 2011