Clinical characteristics and longitudinal associations with obsessive-compulsive disorder in youth exposed to trauma

Abstract

Childhood trauma exposure is associated with posttraumatic stress symptoms (PTSS) and suicidality, however it is also a risk factor for obsessive-compulsive disorder (OCD) in adults. Research examining the relationship between childhood trauma and OCD in youth is mixed, and there is a dearth of research examining the associations among OCD, PTSS, and suicidality. As a result, conclusions have been drawn from primarily cross-sectional adult samples. No study has examined the clinical characteristics associated with OCD in trauma-exposed youth, nor its associations with PTSS and suicidality over time. To address this gap, the present study used logistic regressions and generalized estimating equations in 2068 trauma-exposed youth aged 8—20 who completed assessments at baseline, 6-month, and 12-month follow-ups. In total, trauma-exposed youth with OCD (n = 222, 10.7 %) were more likely to be female (OR = 0.646), had more severe PTSS (OR = 1.032), and more psychiatric comorbidities (OR = 1.391) compared to trauma-exposed youth without OCD. Interpersonal traumas (OR = 1.549) and bullying (OR = 1.294) were associated with a greater likelihood of having OCD; however, these effects were nonsignificant when adjusting for other mental health symptoms. There was no evidence that OCD was associated with the trajectory of PTSS nor suicidality at 6- and 12-month follow-ups. Trauma-exposed youth with OCD may cross-sectionally have more severe clinical presentations overall, but OCD may not be related to the trajectory of these symptoms over time. Future research is needed to understand the directionality of clinical characteristics associated with pediatric OCD and whether interpersonal traumas convey risk uniquely for OCD or for distress in general.

Highlights

• •This study compared trauma-exposed youth with and without OCD over time.
• •Trauma-exposed youth with OCD were more likely to be female than those without OCD.
• •Trauma-exposed youth with OCD had more severe PTSS than those without OCD.
• •Trauma-exposed youth with OCD had more comorbidities than those without OCD.
• •There was no relation between OCD and posttraumatic stress or suicidality over time.

1. Introduction

Trauma exposure, defined as actual or threatened death, serious injury, or sexual violence [1], is common in youth [2]. Nearly 62 % of youth are exposed to trauma before the age of 17 [3], and youth exposed to trauma and neglect are at increased risk for suicidality [5], [4] and post-traumatic stress disorder (PTSD). A lesser considered outcome of childhood trauma is obsessive-compulsive disorder (OCD). OCD is a debilitating psychiatric condition marked by clinically significant and functionally impairing obsessions and compulsions that affects up to 3 % of youth [6], [7] and is more prevalent in individuals exposed to childhood trauma ([8]; Lochner et al., 2002). OCD symptoms in trauma survivors are theorized to develop and/or be used as a means of coping with the trauma and trauma-related symptoms and to prevent future re-traumatization [9], highlighting the need to better understand OCD as a coping mechanism in trauma-exposed youth. Despite this, little is known about the clinical characteristics associated with OCD in trauma-exposed youth, or how OCD may be associated with mental health outcomes in trauma-exposed youth over time. Accordingly, the present study sought to examine the clinical characteristics associated with OCD and the long-term relationships among OCD, posttraumatic stress symptoms (PTSS), and suicidality among trauma-exposed youth.

1.1. Childhood trauma as a risk factor for OCD

Rates of reported trauma exposure among individuals with OCD range widely, between 30 % and 82 % [10], [11], [12], and cumulative trauma exposure is associated with increased risk for OCD [13]. Over three-quarters of adults with OCD report experiencing childhood trauma [14] and sometimes identify traumatic or stressful life events, typically during childhood, as the catatlyst for the development of their OCD [15]. Childhood emotional, physical, sexual abuse, and neglect are associated with greater OCD severity and more severe co-occurring mental health concerns [16], [17], [18], [19], [20], [21], [22], [23], an effect that is particularly strong in pre-pubescent girls [17].

Although there is strong cross-sectional evidence to suggest a relationship between childhood trauma and OCD/obsessive-compulsive symptoms (OCS) in adults, the association between childhood trauma and OCD/OCS in youth is more variable, perhaps because OCS that onset during childhood tend to have a stronger genetic load ([24]; for a review of OCD/OCS in trauma-exposed youth, see [25]). Whereas several studies have found that youth exposed to trauma may evidence higher rates of pediatric OCD [17], [26], [27], [28], [29], [30], and that OCD is associated with greater exposure to childhood trauma [8], [22], [31], other studies have found that trauma confers higher risk for mental health concerns broadly [32], [33], [34]. In addition, many studies use narrow assessments of trauma, oftentimes focusing only on interpersonal traumas (e.g., sexual abuse; [19]), so the broader impact of exposure to a variety of trauma types on OCD development is unknown. While not necessarily meeting the strict definition of a trauma in all cases, bullying can include threat of serious injury or death and is a stressful life experience that has also demonstrated cross-sectional associations with OCD in youth [35], [36], [37].

1.2. OCD and PTSD

Both trauma exposure and PTSD are associated with increased severity and chronicity of OCD symptoms in adults cross-sectionally and over two-, four-, and seven- year follow-ups [38], [15], [39], [40], [14], [41], suggesting that early experiences of trauma modulate the course of OCD over time. Of note, research has yet to examine the reverse – the longitudinal associations between OCD and PTSD/PTSS. Cross-sectionally and among adults, comorbid OCD and PTSD is associated with more severe OCD compared to adults with OCD and no PTSD, however severity of PTSD was similar in adults with comorbid OCD and PTSD compared to adults with PTSD and no OCD [39], [40]. These findings suggest that, cross-sectionally and in adults, the addition of PTSD makes OCD more severe, yet OCD does not appear to impact the severity of PTSD. As OCD that onsets subsequent to PTSD has also been associated with worsened clinical presentation [11], findings altogether point to the development or worsening of OCD in response to trauma (e.g., as compulsions develop to cope with PTSD symptoms or prevent future trauma; [42]).

However, it is not clear whether this relationship may change over time, or, importantly, how it may interact in youth, a noteworthy gap given important developmental considerations in how PTSS presents in youth (e.g., more frequent engagement in risk-taking behavior, interruptions of normal developmental processes; [1], [43]). These considerations are crucial given the shared clinical features of OCD and PTSD, including intrusive and unwanted thoughts and avoidance behaviors, that can further complicate a comorbid clinical presentation [44]. In the absence of research on OCD and PTSS in youth, it is unclear whether and to what extent OCD may relate to PTSS over time.

1.3. OCD and suicidality

Consistent with the literature on OCD prevalence in trauma-exposed youth, there is a dearth of research on suicidality in trauma-exposed youth with OCD, as conclusions are primarily drawn from adults retrospectively reporting on childhood trauma. Interestingly, while OCD severity does not appear to impact suicidality in youth [45], history of childhood trauma does appear to be a critical contributing factor in adult and youth OCD samples[16], [17], [20], [21]. Among adults with OCD, one study found that 82 % of patients who had attempted suicide reported childhood maltreatment compared to 48 % of patients who did not attempt suicide [21]. In this sample, patients who had attempted suicide were likely to have experienced childhood emotional neglect (56 %), physical neglect (54 %), emotional abuse (54 %), sexual abuse (44 %), and physical abuse (48 %), with rates of physical neglect and sexual abuse significantly greater in patients with OCD who had attempted suicide compared to those who had not [21].

The association between OCD and suicidality in trauma-exposed youth is less clear. For example, sexual abuse and OCD unacceptable thoughts (e.g., themes related to sex or harm) are both uniquely associated with suicidal ideation over and above gender, OCD severity, depression, and anxiety, however findings were derived from a mixed age sample (15–80 years old) so it is unclear whether this association holds for youth [20]. In a pediatric sample [17], although trauma exposure and OCS were independently associated with increased suicidal ideation, a nonsignificant interaction between these effects suggested that the experience of trauma did not affect the relationship between OCD and suicidal ideation. However, OCD in this study was assessed using a screener lacking psychometric validation and findings were based on cross-sectional data, so it is not known whether the presence of OCD in trauma-exposed youth may be associated with changes in suicidality over time. Other studies examining suicidality in youth with OCD did not examine the joint effects of trauma exposure and OCD so conclusions regarding these associations cannot be made (e.g., [19]).

1.4. The present study

Many studies establishing a link between trauma exposure and OCD sampled adults who retrospectively reported on childhood trauma (e.g., [16], [18], [21], [23]), or used a combined age sample where the association could be directly tested in youth (e.g., [20], [22]). This represents a notable gap because the relationship between trauma exposure and OCD may differ in youth compared to adults, where youth onset of OCD may be more heavily influenced by genetics [46], [24]. Existing studies on trauma-exposed youth with OCD have yielded inconsistent findings, potentially a result of differences in the assessment of constructs of interest (e.g., trauma and OCD, [17]). More research on the clinical characteristics associated with OCD among trauma-exposed youth is needed. Moreover, no studies across the youth or adult literature have examined the longitudinal associations among OCD, PTSS, and suicidality. Thus, longitudinal research examining how the presence of OCD may relate to PTSS and suicidality over time can further clarify the complex relationship between trauma exposure and OCD in youth.

In a large sample of trauma-exposed youth, we predicted that: (1) the presence of OCD would be associated with older age, female sex, greater exposure to trauma overall, more severe PTSS, and more psychiatric comorbidities at baseline [17], [38], [39], [40], [14], [13]; (2) interpersonal traumas, specifically, would be more strongly associated with the presence of OCD than would non-interpersonal traumas at baseline [16], [17], [18], [19], [20], [21], [22], [23]; and (3) trauma-exposed youth with OCD, compared to trauma-exposed youth without OCD, would evidence more severe and chronic PTSS and suicidality over time [47], [20], [21].

2. Methods

2.1. Procedure

This investigation is based on data collected as part of the Texas Childhood Trauma Research Network (TX-CTRN), a state-wide, multi-center longitudinal study of youth who have experienced at least one lifetime trauma. The TX-CTRN was established in 2020 to develop a state-wide registry of youth, ages 8–20 years, who had experienced a traumatic or stressful life event. This registry facilitates the analysis of outcomes related to trajectories of mental health following trauma and supports development of statistical and longitudinal models of short- and long-term risks and resilience. The network uses a “hub-and-spoke” organization. The hub represents the anchor site where research plans are developed and monitored, and training, outreach, and support is provided to the other sites. The network consists of 12 academic medical center sites (spokes) in Texas that recruit participants from multiple settings, including hospitals, emergency departments, mental health inpatient and outpatient clinics, and primary care clinics.

Informed consent was obtained from the participant, or the parent/guardian if the participant was below the age of 18 years. Youth below the age of 18 years provided assent. Participants were assessed at baseline using the standardized interviews and self-report clinical instruments. Parents/guardians also participated in the interviews when youth were under 18 years old. Follow-up assessments are then conducted at 1- month, 6-months, 12-months, 18-months, and 24-months. The data reported here were obtained on participants enrolled (baseline) between October 3, 2020, and October 10, 2023, and represent findings for those participants who had data up to 12-months. As recruitment and follow-up assessments are ongoing at the time of this report, data from visits beyond 12 months are not included given sample size constraints. Participant compensation consisted of $100 for the baseline visit, $50 for the 6-month visit, and $100 for the 12-month visit. This study was approved by the University of Texas Southwestern Medical Center Institutional Review Board.

2.2. Sample

Participants were recruited from the 12 academic medical center sites comprising the TX-CTRN: University of Texas (UT) Southwestern Medical Center, Baylor College of Medicine, UT-Rio Grande Valley, UT-Health Science Center at Tyler, UT-Health Science Center at Houston, Texas Tech University Health Science Center Lubbock, UT-Austin Dell Medical School, UT-Health Science Center at San Antonio, University of North Texas Health Science Center, Texas A&M Health Science Center, Texas Tech University Health Science Center El Paso, UT-Medical Branch at Galveston. Inclusion criteria include: 1) youth have a history of trauma exposure that fulfills DSM-5 Criterion-A for PTSD, and/or have experienced bullying; 2) between 8 and 20 years of age; 3) able to provide consent for 18–20 year-olds, assent for those under 18 with parental consent; 4) able to speak and read English or Spanish; and 5) function at a level that allows accurate completion of the assessment instruments based on the clinical judgment of the evaluator. Exclusion criteria included either the child or the consenting parent being: a) intoxicated; b) not alert/oriented/or coherent/or medically unstable; c) exhibiting active psychotic symptoms; or d) the child had pervasive developmental disorder/moderate intellectual disability.

2.3. Measures

Concise Health Risk Tracking-Self Report (CHRT-SR-9; [48], [49]). The CHRT-SR-9 [49] is a 9-item self-report measure of suicidality pared down from the CHRT-SR-16 [48]. It uses a Likert-type ordinal scale with item-level responses ranging from 0 (“Strongly Disagree”) to 4 (“Strongly Agree”) to each item when considering the past week (e.g., “I feel that there is no reason to live”). Total score ranges from 0 to 36 with higher CHRT-SR-9 scores indicating greater suicidality risk and higher frequency of suicidal thoughts. To ensure accuracy of the Spanish-translated version of the CHRT-SR-9, it was back-translated by a competent translator and reviewed by a team of Spanish-speaking individuals to reach a consensus on the final translation. The CHRT-SR-9 has an administration time of 8 min. In the present study, the CHRT-SR-9 evidenced strong internal consistency in both English (α = 0.91) and Spanish formats (α = 0.90).

Clinician-Administered PTSD Scale for DSM-5, Child/Adolescent Version (CAPS-CA-5;[50]). The CAPS-CA-5 is a 30-item rater-administered semi-structured interview to measure severity of PTSD symptoms based upon DSM-5 criteria for children and adolescents aged 7 and above. It was modified from the CAPS-5 version for adults to include age-appropriate items and picture response options. Similar to the CAPS-5, the CAPS-5-CA assesses the 20 DSM-5 PTSD symptoms (e.g., “In the past month, have you had any bad dreams about the bad thing that happened or other bad dreams?”). For each symptom, standardized questions and probes are provided. In addition, questions target the onset and duration of symptoms, subjective distress, impact of symptoms on social functioning, impairment in development, overall response validity, and overall PTSD severity. The CAPS-CA-5 was translated into Spanish using the University of Texas Rio Grande Valley Translation and Interpreting Office by a competent translator in both English and Spanish. The CAPS-5-CA has an administration time of 15–50 min. In the present study, the CAPS-CA-5 evidenced strong internal consistency in both English (α = 0.91) and Spanish formats (α = 0.93).

Family History of Depression or Anxiety. Raters asked about family history of depression and anxiety with the parent (if youth is under the age of 18) or directly with the participant if 18 years of age or older (i.e., “Which members of your/your child’s family have been diagnosed with or have experienced the following problems?”). Family history was coded as present if the parent/participant indicated that depression and/or anxiety disorders were present in either of the child’s parents or siblings, and absent if denied.

MINI International Neuropsychiatric Interview for Children and Adolescents English Version 7.0.2 for DSM-5 (MINI-KID;[51]). The MINI-KID is a well-established structured diagnostic interview used to assess the presence of a broad range of DSM-5 child and adolescent psychiatric disorders and risk of suicide. The instrument uses one to four screening questions for each disorder (e.g., “At any time in your life, did you feel sad or depressed?”). If the screening questions are endorsed, additional symptom questions are given for the particular disorder. MINI-KID disorder classifications have shown test-retest reliability and validity comparable to other standardized diagnostic interviews. The MINI-KID has an administration time of 15–50 min. The MINI-KID was used to diagnose OCD and characterize the comorbidities of the sample. A sum comorbidities score was created comprised of the following disorders coded as present = 1 and absent = 0: Major Depressive Disorder, Bipolar I Disorder, Bipolar II Disorder, Other Bipolar Disorder, Panic Disorder, Agoraphobia, Separation Anxiety Disorder, Social Anxiety Disorder, Specific Phobia, PTSD, Alcohol Use Disorder, Substance Use Disorder, Conduct Disorder, Oppositional Defiant Disorder, Any Psychotic Disorder, Bipolar I with Psychotic Features, Anorexia Nervosa, Bulimia Nervosa, Binge Eating Disorder, Generalized Anxiety Disorder, and Adjustment Disorder. All disorders included in the present study were measured at baseline so information regarding relative onset is not available.

Traumatic Events Screening Inventory – Child (TESI-C;[52]): The TESI-C is a 21-item rater-administered interview that assesses a child’s experience of potentially traumatic events, including injuries, hospitalizations, domestic violence, community violence, disasters, accidents, physical and sexual abuse. Two additional items were added for bullying and cyberbullying that were not included in the original TESI-C scale.1 The TESI-C has an administration time of 10–25 min. Consistent with factor analytic work conducted using TX-CTRN data [53], TESI-C items were scored as: Unintentional trauma exposure (directly experienced; e.g., being involved in an accident), intentional interpersonal trauma exposure (directly experienced; e.g., sexual abuse), trauma exposure within the home (witnessed; e.g., witnessing domestic violence), trauma exposure within the community (witnessed; e.g., witnessing physical assault in the community), and bullying/cyberbullying exposure (directly experienced; e.g., being harassed in person or online).

All five instruments were administered by trained research assistants. Shahidullah et al. [54] described the training and certification process for research assistants across the clinical sites. The interrater reliabilities were excellent (see [54]).

3. Data analysis plan

The first aim of this study was to identify baseline clinical characteristics associated with having an OCD diagnosis in trauma-exposed youth. This analysis included 2068 youth; an additional 80 participants were enrolled in the study but were not included in this analysis due to missing age and/or sex. A logistic regression was conducted with OCD diagnosis as the dependent variable and included the following independent variables: age, sex (dummy-coded male vs. other), race (due to cell sizes, these were dummy-coded as White vs. other and Black/African American vs. other [American Indian/Alaskan, Asian, Hawaiian/Pacific Islander, and other]), Hispanic or Latin/e ethnicity, family history of anxiety, family history of depression, trauma exposure, total number of psychiatric comorbidities (excluding OCD), and CAPS-CA-5 severity.

To measure trauma exposure, domain-specific counts of traumatic experiences based on the TESI-C were used across five domains: unintentional experienced (e.g., car accidents, hospitalizations, natural disasters), intentional interpersonal directly experienced (e.g., experiences of abuse, violence, or sexual assault), bullying (inclusive of cyber-bullying and face-to-face bullying), intentional witnessed at home (e.g., witnessing domestic violence), and intentional witnessed in the community (e.g., witnessing community violence). This scoring scheme has shown strong concurrent predictive validity of psychopathology in prior work in this network [53]. A separate logistic regression was also conducted only including these TESI-based trauma categories without adjusting for other covariates to align results with prior work. There was no evidence for multicollinearity across TESI-C trauma categories (rs = 0.13–0.42). Multiple imputation using chained equations [55] was used to estimate missing baseline data, wherein predictors were other baseline variables.2

The final aim was to evaluate the associations between OCD diagnosis at baseline and PTSS severity (CAPS-CA-5) and suicidality (CHRT-SR-9) at 6- and 12-month follow-up assessments. These analyses included 1113 participants with CAPS-CA-5 data (n = 1023 at 6-month and n = 733 at 12-month), and 1162 participants with CHRT-SR-9 data (n = 1067 at 6-month and n = 761 at 12-month). Generalized estimating equations (GEE; [56]) were used to model longitudinal data. Covariates included age, sex, race, Hispanic or Latin/e ethnicity, family history of depression, family history of anxiety, the number of psychiatric comorbidities at baseline, CAPS-CA-5/CHRT-SR-9 scores at baseline, and the wave of data collection, to isolate the associations between baseline OCD and PTSS severity and suicidality over time. To examine differences in the relationships between OCD diagnosis and CAPS-CA-5 and CHRT-SR-9 scores at each timepoint, the 6-month effect, the 12-month effect, and their interaction were also modeled separately to estimate effects at these specific timepoints while adjusting for covariates.3 This analysis used all participants with at least one follow-up visit completed (n = 1169). Missing data were not imputed for the longitudinal response values (although imputed baseline predictors were employed), as some data were missing structurally (i.e., participants had not reached this point in the ongoing study) whereas others reflected dropouts. Patterns of missing data at 6- and 12-month follow-up were examined and did not differ between participants with and without OCD. Imputation, logistic regression, and GEE modeling were performed in Stata v.18 [57].

4. Results

4.1. Demographic characteristics

Of the 2068 participants included in the cross-sectional analysis of baseline clinical characteristics, 222 met diagnostic criteria for OCD (10.7 %) and 1846 (89.3 %) did not. See Table 1 for summary of participant characteristics.

Table 1.

Participant characteristics at baseline.

Variable	No Current OCD Dx	Current OCD Dx	Total
% (n)	89.3 % (1846)	10.7 % (222)	2068
Age in years, mean (SD)	15.1 (3.4)	16.0 (2.9)	15.2 (3.4)
Sex, % (n)
Female	56.0 % (1033)	79.7 % (177)	58.5 % (1210)
Male	44.0 % (813)	20.3 % (45)	41.5 % (858)
Race, % (n)
White	65.2 % (1204)	72.5 % (161)	66.0 % (1365)
Black/African American	15.5 % (286)	8.6 % (19)	66.0 % (1365)
Other	19.3 % (356)	18.9 % (42)	19.2 % (398)
Hispanic or Latin/e, % (n)
Yes	47.1 % (870)	47.3 % (105)	47.1 % (975)
No	52.9 % (976)	52.7 % (117)	52.9 % (1093)
Family History of Depression, % (n)
Yes	51.2 % (646)	67.9 % (91)	52.8 % (737)
No	48.8 % (615)	32.1 % (43)	47.2 % (658)
Family History of Anxiety, % (n)
Yes	47.3 % (596)	61.2 % (82)	48.6 % (678)
No	52.7 % (665)	38.8 % (52)	51.4 % (717)
Unintentional Trauma Exposure(s), % (n)
0 exposures	7.4 % (130)	13.2 % (27)	8.0 % (157)
1 exposure	26.4 % (462)	20.5 % (42)	25.7 % (547)
2 exposures	28.4 % (497)	24.4 % (50)	27.9 % (547)
3 exposures	23.0 % (403)	16.6 % (34)	22.3 % (437)
4 exposures	10.3 % (180)	15.1 % (31)	10.8 % (211)
5 exposures	3.7 % (64)	7.8 % (16)	4.1 % (80)
6 exposures	0.7 % (13)	2.0 % (4)	0.9 % (17)
7 exposures	0.2 % (4)	0.5 % (1)	0.3 % (5)
Intentional Trauma Exposure(s), % (n)
0 exposures	53.3 % (941)	21.1 % (44)	49.9 % (985)
1 exposure	24.4 % (431)	34.9 % (73)	25.5 % (504)
2 exposures	14.3 % (252)	25.4 % (53)	15.5 % (305)
3 exposures	6.7 % (118)	11.0 % (23)	7.1 % (141)
4 exposures	1.2 % (22)	6.7 % (14)	1.8 % (36)
5 exposures	0.1 % (1)	1.0 % (2)	0.2 % (3)
Trauma Exposure w/in the Home, % (n)
0 exposures	64.0 % (1141)	42.8 % (89)	61.8 % (1230)
1 exposure	21.0 % (374)	35.1 % (73)	22.5 % (447)
2 exposures	10.8 % (193)	14.4 % (30)	11.2 % (223)
3 exposures	4.2 % (74)	7.7 % (16)	4.5 % (90)
Trauma Exposure w/in the Community, % (n)
0 exposures	73.0 % (1302)	63.0 % (133)	72.0 % (1435)
1 exposure	20.4 % (363)	23.2 % (49)	20.7 % (412)
2 exposures	5.6 % (99)	11.4 % (24)	6.2 % (123)
3 exposures	1.1 % (19)	2.4 % (5)	1.2 % (24)
Bullying/Cyberbullying Exposure, % (n)
0 exposures	49.4 % (889)	29.2 % (63)	47.2 % (952)
1 exposure	35.2 % (634)	43.1 % (93)	36.0 % (727)
2 exposures	15.4 % (278)	27.8 % (60)	16.8 % (338)
CAPS-CA-5 Severity, mean (SD)	15.1 (12.8)	29.7 (13.8)	16.8 (13.7)
CHRT-SR-9, mean (SD)	9.2 (7.2)	17.4 (8.5)	10.2 (8.2)
Number of comorbidities (MINI-KID), mean (SD)	1.6 (1.9)	4.2 (2.4)	1.8 (2.1)
Comorbidities (MINI-KID), % (n)
Major Depressive Disorder	15.1 % (279)	40.1 % (89)	17.8 % (368)
Bipolar I Disorder	2.3 % (42)	8.6 % (19)	3.0 % (61)
Bipolar II Disorder	0.4 % (7)	0.9 % (2)	0.4 % (9)
Other Bipolar Disorder	0 % (0)	0 % (0)	0 % (0)
Panic Disorder	8.1 % (149)	29.3 % (65)	10.4 % (214)
Agoraphobia	11.1 % (205)	36.0 % (80)	13.8 % (285)
Separation Anxiety Disorder	10.6 % (195)	31.5 % (70)	12.8 % (265)
Social Anxiety Disorder	13.4 (247)	42.8 % (95)	16.5 % (342)
Specific Phobia	13.8 % (254)	28.8 % (64)	15.4 % (318)
PTSD	27.3 % (504)	62.2 % (138)	31.0 % (642)
Alcohol Use Disorder	5.2 % (95)	12.2 % (27)	5.9 % (122)
Substance Use Disorder	9.3 % (171)	20.3 % (45)	10.4 % (216)
Conduct Disorder	4.3 % (80)	10.4 % (23)	5.0 % (103)
Oppositional Defiant Disorder	7.8 % (143)	14.9 % (33)	8.5 % (176)
Any Psychotic Disorder	0.8 % (14)	2.7 % (6)	1.0 % (20)
Bipolar I with Psychotic Features	0.4 % (7)	3.2 % (7)	0.7 % (14)
Anorexia Nervosa	5.3 % (98)	13.5 % (30)	6.2 % (128)
Bulimia Nervosa	2.5 % (46)	8.1 % (18)	3.1 % (64)
Binge Eating Disorder	1.1 % (20)	4.5 % (10)	1.5 % (30)
Generalized Anxiety Disorder	17.0 % (314)	50.0 % (111)	20.6 % (425)
Adjustment Disorder	0.2 % (4)	0.5 % (1)	0.2 % (5)

Note: OCD = obsessive-compulsive disorder; CAPS-CA-5 = Clinician-Administered Posttraumatic Stress Disorder Scale for Children and Adolescents for DSM-5 (range = 0–80); CHRT-SR-9 = Concise Health Risk Tracking Form (range = 0–16); MINI-KID = Mini International Neuropsychiatric Interview for Children and Adolescents; PTSD = posttraumatic stress disorder. Number of comorbidities excludes OCD and is inclusive of the comorbidities listed subsequently.

4.2. Factors associated with OCD diagnosis among trauma-exposed youth

The variables associated with OCD diagnosis at baseline included female sex, OR [SE] = 0.646[0.128], p = 0.027, CAPS-CA-5 severity, OR [SE] = 1.032[0.008], p < 0.001, and number of psychiatric comorbidities, OR [SE] = 1.391[0.059], p < 0.001. No other variables were significantly associated with OCD diagnosis in the model. See Table 2 for complete details.

Table 2.

Clinical characteristics associated with OCD diagnosis at baseline in trauma-exposed youth.

	OR	SE	t	p	95 % CI Lower	95 % CI Upper
Age	0.972	0.027	− 1.04	0.301	0.92	1.03
Male	0.646	0.128	− 2.21	0.027	0.44	0.95
Black/African American	0.620	0.178	− 1.67	0.095	0.35	1.09
Other Race	0.825	0.170	− 0.93	0.351	0.55	1.24
Hispanic or Latin/e	0.792	0.134	− 1.38	0.169	0.57	1.10
Family History of Depression	1.195	0.360	0.59	0.557	0.65	2.20
Family History of Anxiety	1.146	0.315	0.50	0.622	0.66	2.00
Unintentional Trauma Exposures	0.936	0.058	− 1.07	0.285	0.83	1.06
Intentional Trauma Exposures	1.112	0.098	1.20	0.230	0.94	1.32
Trauma Exposure w/in the Home	1.000	0.094	0.01	0.996	0.83	1.20
Trauma Exposures w/in the Community	1.082	0.134	0.64	0.525	0.85	1.38
Bullying/Cyberbullying Exposures	0.850	0.102	− 1.36	0.173	0.67	1.07
CAPS-CA-5 Severity	1.032	0.008	4.12	< 0.001	1.02	1.05
Comorbidities (MINI-KID)	1.391	0.059	7.85	< 0.001	1.28	1.51

Note: CAPS-CA-5 = Clinician-Administered Posttraumatic Stress Disorder Scale for Children and Adolescents for DSM-5 (range = 0–80); MINI-KID = Mini International Neuropsychiatric Interview for Children and Adolescents.

When examining unadjusted effects, experiencing more intentional interpersonal, directly experienced traumas and bullying/cyberbullying experiences were associated with OCD diagnosis. See Table 3 for a summary of unadjusted effects of trauma exposure type on OCD diagnosis.

Table 3.

Traumatic events associated with OCD diagnosis at baseline in trauma-exposed youth.

	OR	SE	t	p	95 % CI Lower	95 % CI Upper
Unintentional Trauma Exposure	0.981	0.055	− 0.34	0.736	0.88	1.10
Intentional Trauma Exposure	1.549	0.112	6.04	< 0.001	1.34	1.79
Trauma Exposure w/in the Home	1.150	0.100	1.60	0.110	0.97	1.36
Trauma Exposure w/in the Community	1.060	0.122	0.51	0.613	0.85	1.33
Bulling/Cyberbullying Exposure	1.294	0.137	2.45	0.014	1.05	1.59

4.3. Association between OCD and PTSS over time among trauma-exposed youth

A total of 1113 participants were included in CAPS-CA-5 longitudinal analyses, including 1023 at month 6 (n = 111, 19.7 % with OCD diagnosis) and 733 at month 12 (n = 82, 19.3 % with OCD diagnosis). Across the 12-month follow-up period, CAPS-CA-5 scores were associated with female sex, B [SE] = -1.882 [0.490], p < 0.001, number of psychiatric comorbidities, B [SE] = .950 [0.203], p < 0.001, and CAPS-CA-5 severity at baseline, B [SE] = .375 [0.0309], p < 0.001. OCD diagnosis at baseline was not significantly associated with PTSS, B [SE] = 1.862 [1.019], p = 0.068. Please see Table 4 for complete details.

Table 4.

Longitudinal association between OCD and posttraumatic stress symptoms in trauma-exposed youth.

	B	SE	t	p	95 % CI Lower	95 % CI Upper
Baseline OCD Diagnosis	1.862	1.019	1.83	0.068	− 0.13	3.86
Age	0.546	0.073	7.46	< 0.001	0.40	0.69
Male	− 1.882	0.490	− 3.84	< 0.001	− 2.84	− 0.92
Black/African American	− 0.351	0.979	− 0.44	0.660	− 1.91	1.21
Other Race	− 0.079	0.611	− 0.13	0.897	− 1.28	1.12
Hispanic or Latin/e	− 0.089	0.516	− 0.17	0.863	− 1.10	0.92
Family History of Depression	0.584	0.683	0.86	0.394	− 0.77	1.93
Family History of Anxiety	0.485	0.776	0.62	0.534	− 1.06	2.03
Comorbidities (MINI-KID)	0.450	0.203	4.67	< 0.001	0.55	1.35
Baseline CAPS-CA-5 Severity	0.375	0.031	12.14	< 0.001	0.31	0.44
Time Point (12-month vs. 6-month)	− 0.096	0.299	− 0.32	0.748	− 0.68	0.49

Note: OCD = obsessive-compulsive disorder; MINI-KID = Mini International Neuropsychiatric Interview for Children and Adolescents; CAPS-CA-5 = Clinician-Administered Posttraumatic Stress Disorder Scale for Children and Adolescents for DSM-5 (range = 0–80). Linear regression model fitted using GEE after multiple imputation.

When examining the effects separately at each timepoint via inclusion of the interaction of baseline OCD diagnosis and follow up time (6 vs 12 months), OCD diagnosis continued to show no significant relationship with CAPS-CA-5 scores at the six-month follow-up, B [SE] = 2.182 [1.133], p = 0.054. The 12-month effect was not significant, B [SE] = 1.381 [1.240], p = 0.265, nor was the interaction term between these two timepoints, B [SE] = -0.800 [1.196], p = 0.503. See Fig. 1 and Supplemental Figs. 1 and 2 for association between OCD diagnosis at baseline on CAPS-CA-5 scores over time.

Fig. 1.

Adjusted trajectory plots of PTSS severity over waves 0, 6, and 12 months based on OCD diagnosis. Note. For each wave, responses are adjusted to the overall mean baseline value of PTSS severity, so that both the overall trend is visible as well as the differences in trends for the two OCD groups. The pink (teal) points are all non-imputed data values for those with (without) OCD at baseline. The light lines connect the points for twenty randomly selected participants for each group to depict the different trajectories realized (without creating an overly dense cloud). The thick lines with the black background are the mean trajectories, and the vertical bars represent the mean + /- SD for each group at each time-point. Only complete data (before imputation) are plotted.

4.4. Association between OCD and suicidality over time among trauma-exposed youth

A total of 1162 participants were included in CHRT-SR-9 longitudinal analyses, including 1067 at month 6 (n = 122, 13.4 % with OCD diagnosis) and 761 at month 12 (n = 90, 11.5 % with OCD diagnosis). Across the 12-month follow-up period, CHRT-SR-9 scores were associated with female sex, B [SE] = -0.719[0.324], p = 0.027, number of psychiatric comorbidities, B [SE] = 0.322 [0.104], p = 0.002, and CHRT-SR-9 severity at baseline, B [SE] = 0.553 [0.0258], p < 0.001. OCD diagnosis at baseline was not significantly associated with suicidality, B [SE] = 0.572 [0.574], p = 0.319. Please see Table 5 for complete details

Table 5.

Longitudinal association between OCD and suicidality in trauma-exposed youth.

	B	SE	t	p	95 % CI Lower	95 % CI Upper
Baseline OCD Diagnosis	0.572	0.574	1.00	0.319	− 0.55	1.70
Age	0.020	0.044	0.46	0.643	− 0.07	0.11
Male	− 0.719	0.324	− 2.22	0.027	− 1.35	− 0.08
Black/African American	0.695	0.490	1.42	0.156	− 0.26	1.65
Other Race	− 0.021	0.412	− 0.05	0.960	− 0.83	0.79
Hispanic or Latin/e	0.212	0.333	0.64	0.524	− 0.44	0.87
Family History of Depression	0.241	0.476	0.51	0.614	− 0.70	1.18
Family History of Anxiety	0.628	0.557	1.13	0.265	− 0.49	1.75
Comorbidities (MINI-KID)	0.322	0.104	3.09	0.002	0.12	0.53
Baseline Suicidality (CHRT-SR-9)	0.553	0.026	21.47	< 0.001	0.50	0.60
Time Point (12-month vs. 6-month).	− 0.481	0.207	− 2.32	0.020	− 0.888	− 0.07

Note: OCD = obsessive-compulsive disorder; MINI-KID = Mini International Neuropsychiatric Interview for Children and Adolescents; CHRT-SR-9 = Concise Health Risk Tracking Form (range = 0–16). Linear regression model fitted after multiple imputation.

When examining the association between OCD diagnosis and CHRT-SR-9 at each timepoint, the interaction term between 6- and 12-month follow-up timepoints was significant, B [SE] = -1.364 [.673], p = 0.043, suggesting that the relationship between OCD diagnosis and change in suicidality differed between these time points, with youth with OCD experiencing less change in suicidality than youth without OCD. That said, when examining the individual effects at each time point, the effect of OCD diagnosis was non-significant at both, 6-month: B [SE] = 1.120 [0.624], p = 0.073; 12-month: B [SE] = -0.244, [0.715], p = 0.733, as their association was in opposite directions.4 See Fig. 2 and Supplemental Figs. 3 and 4 for association between OCD diagnosis at baseline and CHRT-SR-9 scores over time.

Fig. 2.

Adjusted trajectory plots of suicidality severity over waves 0, 6, and 12 months based on OCD diagnosis. Note. For each wave, responses are adjusted to the overall mean baseline value of suicidality severity, so that both the overall trend is visible as well as the differences in trends for the two OCD groups. The pink (teal) points are all non-imputed data values for those with (without) OCD at baseline. The light lines connect the points for twenty randomly selected participants for each group to depict the different trajectories realized (without creating an overly dense cloud). The thick lines with the black background are the mean trajectories, and the vertical bars represent the mean +/- SD for each group at each time-point. Only complete data (before imputation) are plotted.

5. Discussion

The present study sought to examine clinical characteristics cross-sectionally associated with OCD diagnosis in trauma-exposed youth, and the associations among OCD, PTSS, and suicidality in trauma-exposed youth over time. Consistent with hypotheses, OCD diagnosis at baseline was cross-sectionally associated with female sex, greater PTSS severity, and more psychiatric comorbidities, however there was no evidence that OCD diagnosis was significantly associated with age as predicted. In adult samples, the presence of OCD does not significantly impact PTSS severity when compared to individuals with PTSD, however, intuitively, PTSS severity has been found to be greater in adults with co-occurring OCD and PTSD compared to adults with OCD [39], [40]. These earlier findings suggest that the presence of OCD did not make PTSS more severe in adults. In the present sample, however, youth with OCD had more severe PTSS compared to youth who did not have OCD, a robust finding that held even when adjusting for family history, demographic variables, and other psychiatric comorbidities.

There are several potential explanations for this finding. First, differences in OCD and PTSS presentation among youth compared to adults may also translate to differences in the relationships between these symptoms [1], [58], [59], [60], [61], [43], [62]. Second, it is possible that youth in this sample experienced more recent trauma compared to the adults sampled in earlier studies and may therefore be reporting more acute rather than chronic PTSS. It is worth noting that the adult samples compared individuals with probable or diagnosed OCD, PTSD, and co-occurring OCD and PTSD, whereas the present comparisons were only between youth with and without OCD. Thus, it is unclear whether youth with OCD report more severe PTSS compared to youth with a diagnosis of PTSD. Nevertheless, findings suggest that trauma-exposed youth with more severe PTSS are also more likely to meet diagnostic criteria for OCD.

In the unadjusted model examining trauma type only, OCD diagnosis was cross-sectionally associated with experiencing more intentional interpersonal, directly experienced traumas as predicted, as well as with bullying experiences. Intentional interpersonal childhood traumas, such as those involving sexual, physical, or emotional abuse, have been linked to OCD in both youth and adult samples [16], [17], [18], [19], [20], [21], [22], [23]. Our study extends earlier methods by including a wider range of trauma types beyond interpersonal violations, yet findings from the present study still support the relevance of intentional interpersonal traumas above other traumas in pediatric OCD.

Interestingly, being bullied was also cross-sectionally associated with a greater odds of having OCD when examining unadjusted effects. Although not well studied, it has been proposed that peer victimization may lead to the development of OCD through coincidental associations between performing compulsive behavior and teasing-related stimuli (e.g., magical thinking that a ritual performed today will prevent peer victimization tomorrow), or because bullying may contribute to the development of exaggerated cognitions about responsibility or harm [36]. This could explain why victims of severe bullying have a much younger age of onset of anxiety disorders than adults who did not experience bullying during their youth [35]. Conversely, the association may work in the reverse direction – youth with OCD may be subjected to greater bullying by peers due to their symptoms [37]. The null relationship between witnessing trauma in the home (e.g., witnessing parental violence) and OCD is consistent with prior research examining OCD severity in adults [63].

It is important to note, however, that the associations between these trauma types and OCD are only present when examining unadjusted effects and are rendered nonsignificant in the larger model in which demographics, family history, PTSS severity, and comorbidities are included. That these associations between trauma exposure and OCD diagnosis were only present when examining unadjusted effects suggests either that our sample size of youth with OCD was not large enough to detect the unique effects of these trauma types, or more simply that these trauma types may confer risk for psychopathology and distress in general [32], [33], [34]. Future research conducted using a larger sample size of youth with OCD is needed to clearly distinguish whether these trauma types may be uniquely associated with OCD above other psychiatric conditions. Regardless, whether the increased rates of interpersonal traumas among youth with OCD is due to a unique relationship with OCD or to a general risk for psychopathology and distress, our findings suggest that practitioners working with trauma-exposed youth with OCD should be mindful of thoroughly assessing for exposure to and impact of these events.

Next, the present study examined the longitudinal associations among OCD diagnosis, PTSS, and suicidality at 6- and 12-month follow-ups. The relationship between OCD diagnosis at baseline and PTSS over time was not statistically significant overall (p = 0.068), resulting from a borderline nonsignificant effect at month 6 (p = 0.054) and a nonsignificant effect at month 12 (p = 0.26). A nonsignificant interaction effect between month 6 and 12 suggests that relationship between OCD diagnosis and change in PTSS did not differ across timepoints. Similarly, the relationship between OCD diagnosis at baseline and suicidality over time was not statistically significant overall (p = 0.319), resulting from a borderline nonsignificant effect at month 6 (p = 0.073) and a nonsignificant effect at month 12 (p = 0.733). Overall findings in our sample suggest that, contrary to our hypotheses, there was no evidence that OCD diagnosis was associated with PTSS or suicidality over time. While the marginal yet nonsignificant effects on PTSS and suicidality at month 6 cannot be interpreted as evidence of an effect, they do imply the possibility that our analyses were simply underpowered to detect these effects. Future research using larger samples and/or using dimensional rather than categorical assessments of OCD may help elucidate to what extent OCD diagnosis and severity may – or may not – relate to PTSS and suicidality over time. In the absence of replication in a larger sample, however, the present study may indicate that OCD is not a primary driving factor of PTSS and suicidality over time in trauma-exposed youth.

The current study has several limitations to acknowledge. The relatively small sample size of youth with OCD overall (n = 222) and at 6- and 12-month timepoints (n = 122 and 90, respectively) makes it unclear whether marginal and nonsignificant findings suggest no evidence of an effect or whether analyses were simply underpowered. However, confidence intervals were given for all analyses, and these provide further quantitative interpretation of the data because they represent a range of values for parameters which are consistent with the data. For example, the estimated coefficient of 1.86 (95 % CI: − 0.13, 3.86) CAPS-CA-5 severity points comparing the two OCD groups (Table 4) has a lower bound of –0.13, which strongly suggests that a larger sample size might have detected this effect as statistically significant; i.e., with a larger sample size, that lower bound would have been positive. By contrast, the estimated OCD coefficient of 0.57 (95 % CI: − 0.55, 1.70) CHRT-SR-9 severity points is less clear in this regard because the lower bound is quite a bit further away from the reference value of zero. Future similar prospective work should make efforts to design studies that both ensure sufficient balance between the two groups with respect to covariates and have sufficient power to detect such group effects [64]. Additionally, while minor participants are given the option to complete assessments in private, some may choose to complete assessments in front of their parent/guardian. This could potentially impact responding for youth who may be apprehensive to disclose traumatic experiences or mental health symptoms in front of their parent/guardian.

Findings are also limited in that symptom-level data for OCD was not available. Given that there are many shared clinical features between OCD symptoms and PTSS (e.g., intrusive, unwanted cognitions, avoidance of stimuli to control intrusive thoughts; [44]), future studies should assess clinical characteristics on a symptom level to assess the role of symptom overlap on comorbidity. Due to the methodology of the larger study from which this sample was derived, diagnoses on the MINI-KID are available only as dichotomous present/absent. Thus, it is not clear whether OCD symptom severity or presentation may be associated with different risk factors or may uniquely relate to PTSS and suicidality over time. Trauma can impact subthreshold obsessive-compulsive symptoms (e.g., [22]) and different trauma types appear to uniquely map onto different OCD symptoms [65], [15], though these unique relationships may have been missed in the present study. As clinical characteristics were based on cross-sectional data at baseline and longitudinal associations could not account for unmeasured confounds, the directionality of these significant relationships is unclear. It is possible – and perhaps likely – that PTSD/PTSS predate or onset within the same year as OCD [66]. Prospective research following a sample of youth who developed OCD following trauma exposure would provide more precise information regarding risk factors likely to be associated with development of post-traumatic OCD in youth. Lastly, while the scope of this paper was focused on overall trends of PTSS and suicidality, there may be nuance in the symptom trajectories embedded within these overall trends. Future research using samples with more longitudinal data points may consider examining these trajectories more specifically. Such research might also consider exploring how psychological or pharmacological treatment impacts these trajectories. Ongoing work with the TX-CTRN suggests that around half of youth enrolled in the study are currently involved in psychological or pharmacological treatment, and comparison of these groups would allow for a better understanding of the naturalistic course of PTSS and suicidality compared to the course of these symptoms when intervention is provided.

Despite limitations, the present study suggests that trauma-exposed youth with OCD have more severe PTSS and a greater number of psychiatric comorbidities. Trauma-exposed youth with OCD are likely to have experienced more instances of intentional interpersonal directly experienced traumas and bullying, though these experiences may confer risk for distress in general. Lastly, OCD diagnosis does not appear to relate to the trajectory of PTSS and suicidality in trauma-exposed youth over the course of 12 months, suggesting that PTSS and suicidality may be less related to OCD than trauma exposure and other psychosocial factors

Funding

This work was solely supported by the State of Texas through the Texas Child Mental Health Consortium. The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. All authors receive research support from the State of Texas. CMP, JDS, EJB, DBR, KF, and DJN do not receive additional research support from any funding agency. The remaining authors have received research support from the following funding agencies: PJR receives research support from NIH grants (AA029090, MH120599, HD109229); OCO receives research support from NIH grant (3U01MH125062-03S2); EAS receives research funding from The Baylor College of Medicine Intellectual and Developmental Disabilities Research Center (P50HD103555) from the Eunice Kennedy Shriver NICHD. WKG receives research support from NIH (UH3NS100549), IOCDF and McNair Foundation; JFR receives support from NIH grants (LM014306, GM141476, AG077017, MH120599) and an award from NIH division of loan repayment (30TR002103); KDW receives research support from NIH grant (AA029090); CBM receives research support from NIH grants (AA029090, MH117293); AGG receives funding from the Texas Higher Education Coordinating Board, the NIH, and the Misophonia Research Fund; IL from NIH grants (MH113574, TW012463). The contents of this publication do not necessarily reflect the views or policies of the NIH. The mention of trade names, commercial products, or organizations do not imply endorsement by the US Government.

Disclosures

Drs. Pinciotti, Shahidullah, Riddle, Onyeka, Franco, Rousseau, Wagner, and Guzick, and Ms. Bivins have reported no biomedical financial interests or potential conflicts of interest.

PJR serves on a data safety monitoring board for Sunovion Pharmaceuticals.

EAS reports receiving research funding to his institution from the Ream Foundation, International OCD Foundation, and NIH. He was a consultant for Brainsway and Biohaven Pharmaceuticals in the past 12 months. He owns stock less than $5000 in NView (for distribution of the Y-BOCS and CY-BOCS). He receives book royalties from Elsevier, Wiley, Oxford, American Psychological Association, Guildford, Springer, Routledge, and Jessica Kingsley.

WKG receives royalties from Nview, LLC and OCDscales, LLC.

DJN has received research support from Eli Lilly, Glaxo SmithKline (GSK), Janssen, the National Alliance for Research on Schizophrenia and Depression (NARSAD), Navitor, Sage Therapeutics, Takeda Pharmaceuticals, the Texas Health & Human Services Commission, and Wyeth. He has served on speakers’ bureaus and/or received honoraria from Astra-Zeneca, Eli Lilly, GSK, Pfizer and Wyeth. He has served on advisory boards for GSK, Janssen, Merck, and Sage Therapeutics. He has served as a consultant to Sage Therapeutics. Neither he nor family members have ever held equity positions in biomedical or pharmaceutical corporations. CBM, in the last three years, served as a consultant to AbbVie, ANeuroTech (division of Anima BV), Signant Health, Magstim, Inc., Intra-Cellular Therapies, Inc., EMA Wellness, Sage, Silo Pharma, Engrail Therapeutics, Pasithea Therapeutic Corp., EcoR1, GoodCap Pharmaceuticals, Inc., Senseye, Clexio, EmbarkBio, SynapseBio, BioXcel Therapeutics. He is a stockholder with Seattle Genetics, Antares, Inc., Corcept Therapeutics Pharmaceuticals Company, EMA Wellness, Precisement Health, Relmada Therapeutics. He has served on advisory boards for ANeuroTech (division of Anima BV), Brain and Behavior Research Foundation (BBRF), Anxiety and Depression Association of America (ADAA), Skyland Trail, Signant Health, Laureate Institute for Brain Research (LIBR), Inc., Heading Health, Pasithea Therapeutic Corp., Sage. He has served on the Board of Directors for Gratitude America, ADAA, Lucy Scientific Discovery, Inc. He holds the following patents: Method and devices for transdermal delivery of lithium (US 6375,990B1); Method of assessing antidepressant drug therapy via transport inhibition of monoamine neurotransmitters by ex vivo assay (US 7148,027B2).

Appendix A. Supplementary material

Supplementary material