Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 6.
Published in final edited form as: J Trauma Stress. 2020 Oct 1;34(1):56–68. doi: 10.1002/jts.22590

Social–Emotional Profiles of PTSD, Complex PTSD, and Borderline Personality Disorder Among Racially and Ethnically Diverse Young Adults: A Latent Class Analysis

Tanya C Saraiya 1, Skye Fitzpatrick 2, Kathryn Zumberg-Smith 3, Teresa López-Castro 4, Sudie E Back 1,5, Denise A Hien 6
PMCID: PMC9817088  NIHMSID: NIHMS1855158  PMID: 33006199

Abstract

The debate around the construct validity of complex posttraumatic stress disorder (CPTSD) has begun to examine whether CPTSD diverges from posttraumatic stress disorder (PTSD) when it co-occurs with the diagnosis of borderline personality disorder (BPD). The present study (a) examined the construct validity of CPTSD through a latent class analysis of a non–treatment-seeking sample of young trauma-exposed adults and (b) characterized each class in terms of trauma characteristics, social emotions (e.g., shame, guilt, blame), and interpersonal functioning. A total of 23 dichotomized survey items were chosen to represent the symptoms of PTSD, CPTSD, and BPD and administered to 197 trauma-exposed participants. Fit statistics compared models with 2–4 latent classes. The four-class model showed the best fit statistics and clinical interpretability. Classes included a “high PTSD+CPTSD+BPD” class, characterized by high-level endorsement of all symptoms for the three diagnoses; a “moderate PTSD+CPTSD+BPD” class, characterized by endorsement of some symptoms across all three diagnoses; a “PTSD” class, characterized by endorsement of the ICD-11 PTSD criteria; and a “healthy” class, characterized by low symptom endorsement overall. Pairwise comparisons showed individuals in the high PTSD+CPTSD+BPD class to have the highest levels of psychological distress, traumatic event history, adverse childhood experiences, and PTSD symptoms. Shame was the only social emotion to significantly differ between the classes, p = .002, η2 = .16. The findings diverge from the literature, indicating an overlap of PTSD, CPTSD, and BPD symptoms in a non–treatment-seeking community sample. Further, shame may be a central emotion that differentiates between presentation severities following trauma exposure.

There has been considerable debate over the construct validity of complex posttraumatic stress disorder (CPTSD), a trauma- and stressor-related disorder added to the most recent revision of the International Classification of Diseases-11 (ICD-11) in 2018 (World Health Organization [WHO], 2018). Several studies have identified CPTSD as a distinct psychological construct with symptoms divergent from the ICD-11 posttraumatic stress disorder (PTSD) diagnosis (Brewin et al., 2017; Cloitre et al., 2013; Herman, 1992; McBride et al., 2019), whereas, in contrast, others have found evidence suggesting a considerable overlap between the two diagnoses (Landy et al., 2015; Resick et al., 2012; Wolf et al., 2014). Adding to this debate are additional findings that CPTSD may overlap with the comorbidity of PTSD and the Diagnostic and Statistics Manual (fifth ed.; DSM-5) criteria for borderline personality disorder (BPD; American Psychiatric Association [APA], 2013; Jowett et al., 2019; Landy et al., 2015; Resick et al., 2012), propelling discussion of whether the CPTSD construct is unique from the co-occurrence of ICD-11 or DSM-5 PTSD and DSM-5 BPD. The present study aimed to add to the existing literature by (a) completing a latent class analysis (LCA) on an urban, racially and ethnically diverse community sample of trauma-exposed young adults to assess whether the ICD-11 criteria for PTSD and CPTSD and the DSM-5 criteria for BPD would separate individuals into distinct classes and (b) to characterize the classes that emerged with regard trauma characteristics, interpersonal functioning, and social emotions. Investigating the distinguishability of CPTSD among this sample is particularly necessary given that compared to individuals with PTSD, those with CPTSD are more likely to report non-white race/ethnicity (Cloitre et al., 2019), being single, having a low socioeconomic status, and having achieved lower educational attainment (Perkonigg et al., 2016). Moreover, urban, racially and ethnically diverse populations have been shown to have a higher risk of both trauma exposure and PTSD (Gillespie et al., 2009), and all three disorders are associated with young adult populations (Hyland et al., 2019). Accordingly, investigating these three disorders in a racially and ethnically diverse urban sample will provide valuable information.

Disparate classification of CPTSD in diagnostic manuals is one reason there remains a lack of consensus among the field. The ICD-11 included CPTSD as a separate disorder under the category of disorders of extreme stress, whereas the DSM-5 did not, instead broadening the criteria for PTSD to capture the symptoms of CPTSD as well as PTSD (APA, 2013). In the ICD-11, both diagnoses require exposure to a traumatic event followed by overlapping symptom criteria. The PTSD criteria include three symptom groups: reexperiencing the trauma, avoidance of trauma reminders and the trauma itself, and hypervigilance of possible threats in the environment. The CPTSD criteria include the symptoms of PTSD as well as three additional symptom groups that fall under the category of “disorders of self-organization” (DSO). These include affective dysregulation, negative self-concept, and interpersonal disturbances (WHO, 2018). The symptom composition of CPTSD, first theorized by Herman (1992), has been supported by confirmatory factor analyses: CPTSD has additional symptoms beyond PTSD that include disturbances in emotion regulation and interpersonal functioning, and these symptoms are often associated with severe, repeated trauma exposure (Böttche et al., 2018; Brewin et al., 2017; Hyland et al., 2017). Additional work has utilized latent class or latent profile analyses to demonstrate the construct validity of CPTSD in other populations. Most of these studies have found one class with primarily CPTSD symptoms and another class with PTSD symptoms (Böttche et al., 2018; Brewin et al., 2017; Knefel et al., 2015; Perkonigg et al., 2016; Zerach et al., 2019). However, most of these studies have relied on trauma-exposed, treatment-seeking samples, many composed predominantly of women, thereby limiting the generalizability of these findings. The findings from comparable work on trauma-exposed veteran and community samples have not shown CPTSD and PTSD to separate into two classes (Wolf et al., 2014). Indeed, Wolf et al. (2014) found the samples to differ in the degree of symptom severity for both PTSD and CPTSD rather than in symptom composition. This suggests that the construct validity of CPTSD as distinct from PTSD may falter in other, more diverse samples.

Complicating this debate is the speculation of whether CPTSD reflects symptoms falling under the DSM-5 diagnosis of BPD and therefore conceptually capturing the co-occurrence of BPD and PTSD (Jowett et al., 2019; Knefel et al., 2016; Resick et al., 2012). The diagnostic criteria for both BPD and CPTSD include deficits in interpersonal functioning, emotion regulation, and self-perception (APA, 2013; WHO, 2018). Three studies have employed LCA to assess the overlap in or differentiation between these diagnoses (Cloitre et al., 2014; Frost et al., 2018; Jowett et al., 2019), and most have found distinct CPTSD and BPD classes, with some overlap in symptom composition. For instance, in a female, treatment-seeking, trauma-exposed sample, Cloitre et al. (2014) found classes classified as low PTSD; high PTSD; high CPTSD and PTSD symptoms (i.e., CPTSD class); and high symptoms of PTSD, CPTSD, and BPD (i.e., BPD class). The authors concluded that this class structure demonstrated CPTSD to be distinct from BPD. Contradictory to these findings, a recent study extended these findings in a sample of trauma-exposed, treatment-seeking men and women and found that CPTSD and BPD did not separate into unique classes (Jowett et al., 2019). Three classes emerged, classified as CPTSD with high PTSD, CPTSD with moderate BPD and high PTSD, and low PTSD and low BPD (Jowett et al., 2019), suggesting a considerable overlap between CPTSD and BPD. Indeed, even Cloitre et al. (2013) reported that 33.7% of individuals with CPTSD showed a co-occurring diagnosis of BPD. Some researchers have gone further to disentangle this overlap and found that BPD is better characterized by an unstable sense of self and chaotic relationships (Cloitre et al., 2014), whereas CPTSD is characterized by a constant negative sense of self and avoidance of relationships (Brewin et al., 2017; Hyland et al., 2019). However, to continue to parse out these minute differences, additional research is needed on the distinguishability of ICD-11 CPTSD from DSM-5 BPD among other samples. Minimal research has examined whether these constructs uphold in a non–treatment-seeking, racially and ethnically diverse sample of young adults, an examination which may be particularly important given that previous research has shown that a larger proportion of participants categorized in CPTSD classes report being non-white (Cloitre et al., 2019), single, having a lower socioeconomic status, and having achieved less educational attainment (Perkonigg et al., 2016). Consequently, the present study tested the generalizability of the CPTSD construct among a sample that may show a particular vulnerability to CPTSD.

In addition to examining symptom overlap across the constructs of CPTSD, PTSD, and BPD, it is also important to identify their associated characteristics. Deficits in interpersonal functioning (i.e., difficulties in relationships and relating to others) and social emotions (i.e., feelings of shame, guilt, and self-blame) are components of these three disorders in the theoretical and empirical literature. A surge of evidence has increasingly documented the prevalence of shame, guilt, and self-blame across PTSD populations (López-Castro et al., 2019; Pugh et al., 2015) as well as the ways interpersonal dysfunction exacerbates PTSD by increasing social isolation (Monson et al., 2010). The recognition of these qualities in the PTSD literature led to the expansion of the DSM-5 diagnosis to include symptoms related to social emotions and interpersonal relationships. However, these qualities are also descriptive of BPD and CPTSD, as BPD is identified by deficits in interpersonal functioning and chaotic relationships (APA, 2013), and CPTSD is characterized by difficulties in affect regulation and negative self-perceptions (WHO, 2018). Additional research has shown that BPD is also marked by high levels of shame and self-blame (Peters & Geiger, 2016), and among theoretical literature, individuals with CPTSD are described as experiencing pervasive difficulties in relationship functioning (Herman, 1992). This is somewhat unsurprising given that trauma exposure or neglected childhood environments theoretically disrupt developing interpersonal and emotional systems, which thereby takes a toll on emotions related to self-perception and self-worth, in addition to the abilities to relate to and trust in others (e.g., Cloitre et al., 2009; Herman, 1992). Despite these concepts holding true for each disorder, it remains unclear if these constructs differ in the degree to which they exhibit each social emotion and interpersonal dysfunction. Identifying the variation in severity regarding social emotions and interpersonal functioning could demonstrate the associations among these disorders. Such information could move the dialogue from the validity of CPTSD to the ways CPTSD is related to BPD and PTSD, allowing the field to understand the pathway of symptomatic distress that can develop from trauma exposure and the associations among diagnostic constructs (e.g., Hyland et al., 2019).

The present study aimed to expand upon the extant literature by investigating whether the ICD-11 criteria for PTSD and CPTSD and the DSM-5 criteria for BPD would present as distinct classes in a two-, three-, and four-class LCA on a non–treatment-seeking sample of racially and ethnically diverse young adults (N = 197). As LCA is a data-driven approach, we did not hypothesize how many classes would emerge in the best-fitting model. After choosing the best latent class model, we aimed to compare clinical characteristics, social emotions, and interpersonal functioning across the classes. Such an investigation is particularly needed in an urban-dwelling, young adult sample given that these individuals have been shown to have an increased risk of trauma exposure (Gillespie et al., 2009), which can impact interpersonal and social–emotional functioning (Cloitre et al., 2009). Nevertheless, due to the lack of research in this area and the need to identify a best-fitting model before examining variables of interest, we did not speculate about what differences in characteristics would present between the classes.

Method

Data for the following study were derived from an online survey that was part of a larger experimental study with electroencephalography (EEG; Saraiya et al., 2019). The current study was not planned a priori before the parent study.

Participants and Procedure

Participants were recruited from May 2016 to November 2017 through two types of advertisements posted on the the City College of New York’s SONA-system and online community-based sites (e.g., Craigslist and flyers at Columbia University). Participants were first recruited from the former site and then the latter, leading to a sample with a mean young-adult age (i.e., 23 years of age). Flyers recruited individuals who had experienced trauma exposure and individuals who identified as healthy adults. Interested participants contacted the lab and either completed a phone screen with a research assistant or an online screening tool to assess the general eligibility criteria. Inclusion criteria were: age 18–65 years, English fluency, physically healthy enough to sit at a computer, and self-reported normal or corrected vision. Exclusion criteria were participant self-report of a past or current diagnosis of psychiatric illness from a healthcare professional of bipolar I disorder, schizophrenia, schizoaffective disorder, delusional disorder, and brief psychotic disorder; current use of psychotropic medications; past or current diagnosis of a neurological syndrome (i.e., seizure disorder, brain trauma, and/or tumor disorder); and hairstyles (e.g., braids, weaves, locks) that would obstruct EEG recordings of brainwave activity. Participants with hairstyles obstructing EEG recordings were given the opportunity to return to the study at a later time or complete the study without EEG recordings. For both groups, eligible participants were forwarded to an online consent form, which they signed electronically. After informed consent was received, participants were immediately forwarded to the online survey. Participants from the community received $50 (USD) for completion of the entire study, and participants from the SONA system received $40 and two study credits for completion of the entire study. All procedures were approved by the institutional review boards at The City College of New York and Adelphi University.

Of the 252 participants who met the eligibility criteria and completed the online survey, 204 endorsed exposure to traumatic events on the Life Events Checklist for DSM-5 (LEC-5; Weathers, Blake, et al., 2013). Removal of incomplete data on measures of interest for this latent class analysis (n = 7) led to a total of N = 197 participants in this study. Of these 197 individuals, 71.1% were women, and 28.9% were men. Most individuals were of young adult age (M = 22.94 years, SD = 5.12, range: 17–41 years). In total, 38.1% of the sample reported their race as Black/African American, 28.9% as Caucasian, 24.9% as Asian, 2.5% as Native American/American Indian, and 5.6% self-identified as “other.” Regarding ethnicity, 40.1% reported that they were Hispanic. Participants reported a low socioeconomic status for an urban residence (M = $32,750.92, SD = $26,849.97) and had completed approximately 13.56 years (SD = 1.74) of education, which is equivalent to some college. Most participants were single (91.4%) and had no children (94.9%). The primary language for most participants was English (84.8%). Regarding religion, most participants reported being Christian (41.6%), followed by nonreligious (38.1%), Islamic (10.7%), other (8.6%), and Jewish (1.0%). On average, participants endorsed 3.61 (SD = 2.47) lifetime traumatic experiences on the LEC-5. Intentional traumas (see the Measures section) were the most common (M = 1.38, SD = 1.31), followed by accidental traumas (M = 1.25, SD = 1.06) and other traumatic events (M = 0.99, SD = 0.95). On average, participants reported a level of PTSD symptoms that was slightly under the established diagnostic cut off 33 for the PTSD Checklist for DSM-5 (PCL-5; Weathers, Litz, et al., 2013; M = 26.55, SD = 20.48).

Measures

Demographic Characteristics

Age, gender, race/ethnicity, and years of education were assessed using basic demographic questions. Participants self-identified their race and ethnicity and were categorized in the following way: White, Black or African American, Native American, Asian, or Other. Participants identified their ethnicity as Hispanic or non-Hispanic.

Trauma Exposure

The LEC-5 (Weathers, Blake, et al., 2013) was used to assess exposure to traumatic events as defined by the DSM-5 (APA, 2013). Responses were dichotomized (1 = “yes” or 0 = “no”) and summed to provide a total number of trauma exposures for each participant. Three categories of trauma exposure were created based on LEC item endorsement: accidental traumas (e.g., natural disaster, fire/explosion, transportation accident, serious accident at work or home, and exposure to a toxic substance), intentional traumas (e.g., physical assault, sexual assault, other unwanted sexual experience, combat, and captivity), and other traumas (e.g., life-threatening illness or injury, severe human suffering, sudden violent death, sudden accidental death, serious injury or harm you caused to someone else, and any other stressful experience). Previous research has found the LEC-5 to have a high interrater agreement (κ = 0.61) among a sample of college students (Gray et al., 2004).

Adverse Childhood Experiences

To assess exposure to childhood adversity (i.e., before 18 years of age), participants completed the 10-item, mini Adverse Childhood Experiences (MINI-ACE) scale (Felitti et al., 1998). Dichotomized responses were summed to give a total score for the number of adverse childhood events each participant experienced. The ACE has been widely used to assess childhood trauma exposure in both high- and low-income countries (Anda et al., 2010).

PTSD Symptoms

Posttraumatic stress symptoms were assessed using the PCL-5, a 20-item self-report measure that corresponds with symptoms from the DSM-5 (APA, 2013; Weathers, Litz, et al., 2013). Selected items from this DSM-5 measure were used to approximate the ICD-11 PTSD diagnosis of PTSD, as has been done in previous work (Cloitre et al., 2014). Respondents rated items using a 5-point Likert-type scale ranging from 0 (not at all) to4 (extremely). Responses were dichotomized for the LCA so that participant responses of 0–1 corresponded to no symptom endorsement, and responses of 2–4 signified symptom endorsement, as suggested by psychometric literature on the PCL-5 (Weathers, Litz, et al., 2013). Overall sum scores (range: 0–80) and sum scores for each cluster were used to assess the degree of PTSD symptom severity. Per the DSM-5 criteria, Cluster B includes intrusion symptoms, Cluster C includes avoidance symptoms, Cluster D includes negative alterations in cognition and mood, and Cluster E includes symptoms related to alterations in arousal and reactivity. The PCL-5 has shown good reliability and validity (McDonald & Calhoun, 2010). In the present sample, the Cronbach’s alpha value was .96.

Psychological Distress

Psychological distress was assessed using the 53-item, self-report Brief Symptom Inventory (BSI; Derogatis, 2001). Participants were asked to endorse how frequently they were distressed by each symptom, rating responses on a scale of 0 (not at all) to 4 (extremely). Responses were dichotomized for the LCA, with a score of 0–1 indicating no endorsement and a score of 2–4 indicating item endorsement, as done in previous analyses (Cloitre et al., 2014). Sum scores were used to create the Global Severity Index (range: 0–212), whereby higher scores indicate more psychological distress. The BSI has shown good psychometric properties (Wiesner et al., 2010). In the present sample, the Cronbach’s alpha value was .98.

BPD Symptoms

Fifteen items from the Structured Clinical Interview for Disorders for the DSM-IV (SCID-II; First & Gibbon, 2004) were included in the parent study to approximate DSM-5 BPD symptoms, and 10 items were specifically selected for the present study to represent the nine DSM-5 BPD symptoms; this corresponds with previous work (e.g., Cloitre et al., 2014). Participants reported “yes” or “no” to symptom endorsement. Previous research has shown the SCID-II to be a reliable measure of personality disorder diagnosis (Jacobsberg et al., 1995). The Cronbach’s alpha value in the present sample was .85.

Social Emotions

The 11-item Tests of Self-Conscious Affect (TOSCA-3S; Tangney et al., 1989) was used to assess three social emotions: shame self-talk, guilt self-talk, and blaming. A total of 86 participants completed this measure, as it was added later to the battery of questionnaires in the parent study. Participants were asked to imagine themselves in 11 different scenarios, and for each scenario, they were asked to respond to three provided reactions. Participants reported how likely they were to exhibit each reaction, using a scale of 1 (not likely) to 5 (very likely). Responses to each reaction are categorized and summed into three subscales—Shame Self-Talk, Guilt Self-Talk, and Blaming—whereby higher scores denote more severity. The TOSCA-3S has demonstrated good validity for both the Shame Self-Talk and Guilt Self-Talk subscales (Luyten et al., 2002). In the present sample, the Cronbach’s alpha value for the total measure was .88.

Interpersonal Functioning

Difficulties in interpersonal functioning and distress generated from interpersonal situations were assessed using the 32-item self-report Brief Inventory of Interpersonal Problems–Circumplex–Item Response Theory (IIP-C-IRT; Sodano & Tracey, 2011). The IIP-C-IRT has two sections. The first section presents items following the prompt, “It’s hard for me to…,” and the second section presents items following the prompt, “The following are things you may do too much…”. Responses range from 0 (not at all) to4 (extremely). Total scores were calculated by taking the mean score across participant responses. Higher scores suggest a higher degree of interpersonal dysfunction. The IIP-C-IRT has demonstrated good internal consistency and validity (Sodano & Tracey, 2011). In the present sample, the Cronbach’s alpha value was .93.

Data Analysis

Items representing PTSD, BPD, and CPTSD symptoms were based on prior research studies, and similar measures and methods were used (e.g., Cloitre et al., 2013, 2014; Jowett et al., 2019). In total, 23 items were used, whereby seven items represented PTSD symptoms, six additional items represented distinct CPTSD symptoms, and 10 items represented BPD symptoms (Supplementary Table S1). When possible, two items were used to represent one symptom to increase the variance in participant responses. Included items were dichotomized and coded as 0 for no symptom endorsement and 1 for symptom endorsement, in alignment with procedures for LCA.

Three latent class models were run in Mplus (Version 8.2) with a maximum likelihood estimation (Muthén & Muthén, 1998). All models were run with 40 initial starts and eight final stage optimizations, in addition to 400 initial starts and 100 final optimizations to replicate the best log-likelihood values. All models converged and replicated the best log-likelihood values at both initial starts and final optimizations. Notably, five- and six-class models were also completed, but the best log-likelihood value was not replicated, and, thus, these models were not interpreted or reported. Three iterative models of two, three, and, four classes were evaluated and compared on the following diagnostics: Akaike information criterion (AIC), Bayesian information criterion (BIC), sample-size–adjusted BIC (ssaBIC), entropy, and the Lo–Mendell–Rubin (LMR) adjusted test. Generally, lower BIC values suggest superior model parsimony, whereas higher entropy values, closest to the value of 1, suggest superior class separation. The LMR likelihood ratio test compares one fewer latent class, Ho = k-1, to the current number of classes in a model, Ha = k, to determine which number of classes better represents the data. Significant values suggest that the current model of classes, Ha = k, is superior to the model with one fewer class. Classes were interpreted based on the estimates for each item. Generally, estimates of approximately 0.50 or higher were interpreted as a high endorsement of an item (Supplementary Table S2). The probabilities for each individual’s latent class membership were saved and used to compare classes with regard to trauma exposure characteristics, social emotions, and interpersonal functioning.

One-way analyses of variance (ANOVAs) were used to test whether classes differed on each clinical variable of interest. Due to the completion of 15 one-way ANOVAs, which increases the possibility of Type I error, alpha levels were adjusted by dividing the standard alpha by the number of tests (i.e., .05/15) to reach a new significance of α = .003. Post hoc Bonferroni tests were used to compare the mean of one class to the mean of all other respective classes with regard to the variable of interest. Significant differences between classes were determined by an alpha level of .05 for post hoc tests. The Bonferroni post hoc test is conservative and adjusts for the familywise error rate of multiple comparisons; thus, the alpha level for post hoc tests was not adjusted. These analyses were completed in SPSS (Version 20.0). In addition, a sensitivity power analysis was completed in G*Power on the subset of participants (n = 86) who completed the TOSCA-3S. A one-way ANOVA with an alpha level of .05, power of .80, and four groups demonstrated an effect size of f = 0.37 for a sample size of 84 (i.e., a multiple of four for four groups). As Cohen’s f is half of Cohen’s d in magnitude, this effect size is considered large. Thus, this suggests that only large differences between groups on the TOSCA-3S were detectable.

Results

LCA

Iterations of two-, three-, and four-class latent class models were fit to the data. The two-class model showed one class (44.7% of the sample) with more endorsement of all PTSD, CPTSD, and BPD symptoms and a second class with minimal symptom endorsement (55.3%). The three-class model showed one class (49.2%) with low endorsement on all items, a second class (14.2%) with high endorsement on the BPD symptoms, and a third class with a high-level endorsement of all PTSD, CPTSD, and BPD symptoms (36.5%). Per our interpretation, the four-class model showed one class with a high endorsement of all symptoms (24.4%), termed the “high PTSD+CPTSD+BPD” class; a second class with a moderate endorsement of all symptoms (17.3%), termed “moderate PTSD+CPTSD+BPD” class; a third class with a high endorsement of PTSD symptoms (13.7%), termed the “PTSD” class; and a fourth class with a low endorsement of all symptoms (44.7%), termed the “healthy” class. Notably, individuals in the second class showed moderate endorsement of four PTSD symptoms and low endorsement of three symptoms (estimate range: 0.38–0.62; see Supplementary Table 2). Given that each of the items with moderate endorsement represented a unique PTSD symptom category (i.e., reexperiencing, avoidance, and hyperarousal), we interpreted this to suggest moderate PTSD endorsement. A comparison of fit statistics and clinical interpretability demonstrated the four-class model to be the best fit to these data (Table 1, Figure 1). Although entropy, the quality of the class specification, was higher for the three-class model than the four-class model, the four-class model showed the lowest BIC in addition to a significant LMR adjusted test. Supplementary Table 2 shows the probability of endorsement for each symptom among each class in each model.

Table 1.

Fit Statistics for Latent Class Models

Model AIC BIC ssaBIC Entropy Adjusted LMRT
2 class 4,745.396 4,899.707 4,750.813 0.938 1,016.98***
3 class 4,583.432 4,816.539 4,591.614 0.958 208.322**
4 class 4,481.786 4,793.690 4,492.735 0.947 148.475*
Open in a new tab

Note. Bolded figures show the final selected model. AIC = Akaike information criterion; BIC = Bayesian information criterion; ss-BIC = sample size–adjusted BIC; LMRT = Lo-Mendell-Rubin test.

*

p < .05.

**

p < .01.

***

p < .001.

Figure 1.

Figure 1

Open in a new tab

Four-Class Model

Note. This graph depicts the best fitting latent class model. On the horizontal axis are items included in the latent class models, with shaded areas showing which items correspond to which diagnoses. The four lines on the graph represent the four classes that emerged from the model; the proportion of symptom endorsement (i.e., estimates) is shown on the vertical axis. PTSD = posttraumatic stress disorder; CPTSD = complex PTSD; BPD = borderline personality disorder.

Social–Emotional Profiles

Comparisons of clinical characteristics, social emotions, and interpersonal functioning among the four classes from the four-class model revealed significant group differences at the p = .003 level (Table 2). Each class significantly differed from one another on psychological distress, post hoc p < .001 to p = .013, ds = 0.98–4.06. Individuals in the high PTSD+CPTSD+BPD class showed the highest level of distress (BSI, M = 102.67, SD = 30.69), followed by those in the moderate PTSD+CPTSD+BPD class (M = 68.59, SD = 27.01), PTSD class (M = 32.67, SD = 21.82), and the healthy class (M = 17.52, SD = 12.94). In comparison to the other classes, participants in the high PTSD+CPTSD+BPD class also showed the highest level of PTSD symptoms (M = 51.92, SD = 11.60), ps < .001, ds = 1.65–3.80); the most reported lifetime traumatic events (M = 5.67, SD = 2.78), p < .001–p = .006, ds = 0.67–1.54; and the most reported childhood adversities (M = 4.81, SD = 2.48), p < .001–p = .014, ds = 0.59–1.61. Regarding symptom clusters, individuals in the PTSD class showed more intrusion symptoms (Cluster B: M = 10.48, SD = 3.04), p < .001, d = 2.02, and avoidance symptoms (Cluster C: M = 5.63, SD = 1.76), p = .014, d = 0.68, than those in the moderate PTSD+CPTSD+BPD class (Cluster B: M = 6.18, SD = 3.17; Cluster C: M = 4.09, SD = 2.67). Further, participants in both the high PTSD+CPTSD+BPD and moderate PTSD+CPTSD+BPD classes showed higher levels of interpersonal dysfunction than those in the PTSD and healthy classes, ps < .01, ds = 0.95–1.78. Concerning social emotions, ratings of guilt self-talk and blaming did not differ between the four classes. However, ratings of shame self-talk were significantly higher for individuals in the high PTSD+CPTSD+BPD class (M = 35.75, SD = 10.52) than those in the PTSD (M = 26.93, SD = 7.56), p = .022, d = 0.90, and healthy classes (M = 27.57, SD = 8.38), p = .009, d = 0.84.

Table 2.

Clinical and Social–Emotional Characteristics of Participants in the Four Class Model (N = 197)

Variable High
PTSD+CPTSD+BPD
(n = 48)
 Moderate
PTSD+CPTSD+BPD
(n = 34)
 PTSD
(n = 27)
 Healthy
(n = 88)
 F or Welch’s F p η2
M SD M SD M SD M SD
LEC-5 sum score 5.67a 2.78 3.58b 2.11 3.96b 1.99 2.40c 1.68 20.63 < .001 .28
 Accidental traumas 1.77a 1.28 1.15b 0.93 1.11b 1.12 1.05b 0.87 4.11 .009 .08
 Intentional traumas 2.40a 1.33 1.41bb 1.08 1.89ab 1.31 0.65c 0.87 26.70 < .001 .31
 Other traumas 1.50a 1.22 1.03a 0.87 0.96a 0.81 0.70b 0.73 6.14 .001 .11
ACE sum score 4.81a 2.48 3.29b 2.67 2.96bc 2.75 1.69c 1.57 22.27 < .001 .25
PCL-5 symptom scores 51.92a 11.60 31.47b 13.46 35.96b 8.33 7.92c 6.27 260.54 < .001 .79
 Cluster B 12.83a 3.39 6.18b 3.17 10.48c 3.04 2.22d 2.27 156.89 < .001 .72
 Cluster C 5.60a 1.89 4.09b 2.67 5.63a 1.76 1.23c 1.68 83.28 < .001 .52
 Cluster D 18.88a 5.72 13.65b 6.61 10.41c 4.82 2.31d 2.57 153.83 < .001 .69
 Cluster E 14.60a 4.25 7.56b 4.49 9.44b 4.14 2.17c 2.44 132.24 < .001 .67
Shame self talka 35.75a 10.52 34.57ac 9.20 26.93bc 7.56 27.57bc 8.38 5.37 .002 .16
Guilt self talka 42.22 8.99 41.36 11.09 44.36 6.39 43.70 6.86 0.43 .73 .02
Blaminga 25.69 9.33 19.86 5.53 19.79 6.14 20.70 7.21 3.51 .019 .11
IIPb 2.94a 0.63 2.61a 0.66 2.03b 0.55 1.93b 0.53 35.88 < .001 .36
BSI 102.67a 30.69 68.59b 27.01 32.67c 21.82 17.52d 12.94 137.03 < .001 .72
Open in a new tab

Note. Mean values with different subscripts in the same row denote significant group differences at α = .05 in post hoc Bonferroni tests. IIP = Inventory of Interpersonal Problems; BSI = Brief Symptom Inventory.

a

Total N = 86; high PTSD+CPTSD+BPD, n = 36; moderate PTSD+CPTSD+BPD n = 14; PTSD, n = 14; healthy, n = 23.

b

Total N = 196; high PTSD+CPTSD+BPD, n = 47; moderate PTSD+CPTSD+BPD n = 34; PTSD, n = 27; healthy n = 88.

Discussion

The present study aimed to assess whether the ICD-11 criteria for PTSD and CPTSD and the DSM-5 criteria for BPD would emerge separately among a sample of racially and ethnically diverse trauma-exposed young adults through a series of LCAs and, secondly, to ascertain the differences in the trauma characteristics, social emotions, and interpersonal functioning among classes in the best-fitting model. This was the first study, to our knowledge, to complete an LCA of PTSD, CPTSD, and BPD among a sample of non–treatment-seeking, racially and ethnically diverse young adults, and, notably, our results diverge from the current literature. The four-class model was superior to the two- and three-class models and included a high PTSD+CPTSD+BPD class, a moderate PTSD+CPTSD+BPD class, a PTSD class, and a healthy class. Individuals in the high PTSD+CPTSD+BPD class showed the poorest overall functioning, followed by those in the moderate PTSD+CPTSD+BPD class, PTSD class, and finally the healthy class. Accordingly, these results suggest that CPTSD and BPD may not distinguish from PTSD or from one another in a non–treatment-seeking, racially and ethnically diverse sample of young adults. Second, these findings show that when PTSD, CPTSD, and BPD co-occur, this constellation of symptoms shows poorer functioning in comparison to PTSD.

Complex PTSD was distinguished from PTSD but not from BPD, which diverges from other LCAs on PTSD, CPTSD, and BPD symptoms (Cloitre et al., 2014; Frost et al., 2018; Jowett et al., 2019). This is notable given that we used a similar methodology to that used by Cloitre et al. (2014), such as which DSM-5 items were used to approximate the ICD-11 criteria for PTSD and CPTSD. However, in the present study, both the three- and four-class models showed CPTSD and BPD to never separate from one another, suggesting that CPTSD and BPD overlap highly in a young adult community sample. This lack of distinction can be interpreted in a few ways. First, it is somewhat logical that CPTSD and BPD overlap given that emotion dysregulation, disrupted identity, and interpersonal difficulties are part of the diagnostic criteria for both disorders. Previous literature has suggested that CPTSD is theoretically distinguishable from BPD because CPTSD is characterized by a more consistent negative sense of self (Brewin et al., 2017), but some research suggests that this also characterizes BPD (Vater et al., 2015), leading to somewhat inconclusive descriptions on the phenomenology of each disorder. What remains clear is that unlike BPD, CPTSD requires an index traumatic event. It should be noted, however, that the expectation that BPD and CPTSD would separate from one another rests on the assumption that BPD is also a discrete diagnosis with limited heterogeneity. This assumption is speculative given that a recent network analysis on PTSD, CPTSD, and BPD demonstrated considerable variance in the BPD diagnosis (Knefel et al., 2016). Thus, our results may simply demonstrate that BPD and CPTSD are two heterogeneous, overlapping disorders and, thus, that they thus present concurrently in some individuals with PTSD, a finding that is emerging in recent literature (e.g., Hyland et al., 2019). Understanding the utility of this overlap will be an important area of investigation for future clinical and research work.

Nevertheless, the lack of differentiation between CPTSD and BPD points to meaningful differences in treatment-seeking and non–treatment-seeking trauma-exposed samples. Although prior studies on treatment-seeking samples have demonstrated three respective classes of PTSD, CPTSD, and BPD (e.g., Cloitre et al., 2014; Jowett et al., 2019), the absence of these findings in the present sample may point to differences in clinical severity. Previous studies in this area have assessed samples in which either all participants endorsed childhood sexual or physical abuse (e.g., Cloitre et al., 2014), lifetime sexual abuse (Frost et al., 2018), or more traumatic events than the current sample (e.g., Jowett et al., 2019). This suggests that the distinction between CPTSD and BPD may be more apparent among severely traumatized individuals and less so in less traumatized samples, as in this study. Put another way, CPTSD may only emerge as a distinct class when there is a sample that reports higher levels of childhood trauma exposure. This could then suggest that differences in clinical presentations following trauma exposure may depend more on the severity of trauma exposure than on diagnostic differences. It may be more pertinent to see additional symptoms of CPTSD and BPD as clinical indicators of the severity of traumatic event histories.

It should be noted that there are two important caveats to this interpretation. First, we anticipated that CPTSD would present as a distinct class in our sample given that compared to PTSD, a larger proportion of participants with CPTSD in prior studies have been non-white (Cloitre et al., 2019), single, living alone, and of lower socioeconomic status (Perkonigg et al., 2016). The absence of this finding could suggest that CPTSD and BPD are more likely to co-occur in samples that are more heterogeneous, younger, and socioeconomically poorer but, at the same time, may be a consequence of the unfortunate evidence that BPD also develops in individuals with lower socioeconomic statuses (Cohen et al., 2008). Symptoms of CPTSD and BPD may thus point to structural inequalities in society that increase the risk of chronic exposure to trauma in addition to living in poverty and, thus, present in tandem with PTSD symptoms rather than separately. Another caveat may be the parameters of the present sample. By focusing on a trauma-exposed sample akin to prior studies, we did not include non–trauma-exposed individuals with primarily BPD symptoms. Thus, the analysis may have thwarted the development of a BPD class separate from the PTSD classes given that the trauma-exposed nature of the sample resulted in spuriously elevated PTSD symptomology across groups. Although this is plausible, Cloitre et al. (2013) completed a latent profile analysis on a trauma-exposed sample of individuals both with and without a subset of participants with BPD and found the same results. Nevertheless, additional LCAs on both trauma-exposed and non–trauma-exposed participants with BPD symptoms is warranted.

The present study also provided unique information on the traumatic event characteristics, social emotions, and interpersonal functioning across the four classes in this study. Individuals in the high PTSD+CPTSD+BPD class exhibited the most psychological distress, PTSD symptoms, and childhood and lifetime trauma exposure compared to the other classes. In addition, participants in both the high PTSD+CPTSD+BPD and moderate PTSD+CPTSD+BPD classes showed higher levels of interpersonal dysfunction and shame than those in the PTSD and healthy classes. These findings are consistent with the broader literature that suggests that BPD coupled with PTSD and CPTSD, respectively, are associated with a higher degree of psychological distress and poorer functioning (Karatzias et al., 2019; Powers et al., 2017; Zerach et al., 2019). This also aligns with theoretical literature and diagnostic criteria that characterize CPTSD and BPD as including negative self-perceptions and interpersonal dysfunction (APA, 2013; Linehan, 1993).

The high rates of lifetime trauma exposure and childhood adversity in the high PTSD+CPTSD+BPD class were somewhat anticipated. Individuals with BPD often exhibit elevated childhood trauma exposure or neglect (Spatz Widom et al., 2009; Turniansky et al., 2019), and such childhood environments have also been linked with the development of CPTSD (e.g., Briere & Spinazzola, 2005; Cloitre et al., 2009; Herman, 1992). Thus, these results support the notion that chronic and early trauma exposure may lead to the interpersonal, emotional, and identity-based dysregulation that characterizes BPD and, by virtue of their overlap, CPTSD as well. However, these findings may also help discern between moderate and severe presentations of PTSD, CPTSD, and BPD. Participants in the high PTSD+CPTSD+BPD class showed more PTSD symptoms, psychological distress, and lifetime trauma exposure than those in the moderate PTSD+CPTSD+BPD class. This information, although requiring replication, may provide preliminary evidence on how severe presentations of combined PTSD, CPTSD, and BPD may logically be linked with more extensive trauma histories.

Finally, comparisons in social emotions across the four classes may assist with distinguishing between these disorders. Individuals in the high PTSD+CPTSD+BPD class reported higher levels of shame in comparison to those in the PTSD and healthy classes; there was no difference in the levels of shame between individuals in the high PTSD+CPTSD+BPD and moderate PTSD+CPTSD+BPD classes. Shame—a pervasive negative appraisal of the self (Tangney, 1995)—may be a defining feature of presentations following trauma exposure that is also characterized by symptoms of PTSD, CPTSD, and BPD. Comparatively, guilt and blame may be more universal responses to traumatic events regardless of symptom presentation. These findings contradict past research that has suggested that individuals with CPTSD and BPD show higher levels of guilt than those with traditional PTSD presentations (Courtois & Ford, 2009), and instead suggests that shame may be a social emotion that can discern between severe PTSD presentations with CPTSD and BPD and presentations that manifest as primarily ICD-11 PTSD symptoms. This also aligns with the literature suggesting that shame is more severe than guilt in psychopathology (Candea & Szentagotai-Tatar, 2018; Peters & Geiger, 2016). Notably, only a subset of participants completed the questionnaire used to assess self-talk, guilt self-talk, and blame. According to the sensitivity power analysis we completed, differences in shame, guilt, and blame were only detectable if there was a large difference (i.e., an f value of 0.37 or higher). Consequently, it is plausible that the differences in guilt and blame were small-to-medium effects that were not detected due to reduced power. Thus, these results should be seen as preliminary.

With regard to clinically attending to individuals with PTSD, CPTSD, and BPD symptoms, the present results suggest that treatments should go beyond PTSD symptom reduction to address emotion dysregulation, interpersonal functioning, and shame-based cognitions. Evidence-based treatments for CPTSD have been developed to build up core elements of interpersonal functioning and help clients regulate and organize their emotional states, such as Trauma Affect Regulation (i.e., TARGET; Ford, 2015) or Skills Training in Affective and Interpersonal Regulation (i.e., STAIR; Cloitre & Schmidt, 2015). These interventions could be combined with treatments that address BPD and PTSD, such as dialectical behavior therapy with prolonged exposure (Harned & Linehan, 2008), and with compassion-based interventions that address shame and the disrupted ability to trust others (Karatzias et al., 2019). Although additional research on how to best address co-occurring symptom presentations is needed, the current findings support the need to combine existing evidence-based approaches to address the multiple converging emotional and behavioral disruptions of the most complex diagnostic profiles that emerge after trauma exposure.

Several limitations of the present study are worth noting. First, all symptoms were measured using self-report instruments. This may have resulted in an increased likelihood of misclassification, as individuals may have endorsed items that a clinician using a structured clinical interview would not. Second, although our assessment of PTSD and BPD were conducted with well-validated DSM-5-based measures, we did not use a validated measure for CPTSD (i.e., the ICD-11 Trauma Questionnaire; Hyland et al., 2017) given that it was published after the parent study began. Rather, consistent with the work by Cloitre et al. (2014), we approximated CPTSD from available measures. Of note, Hyland et al. (2019) and Jowett et al. (2019) used the ICD-11 Trauma Questionnaire and found that CPTSD and BPD retained their overlap, suggesting that even if we used this measure, our results may have produced similar findings. Nevertheless, future studies should use validated CPTSD measures in similar samples to replicate the present study’s findings. Third, this study compared the DSM-5 diagnosis of BPD with the ICD-11 criteria for PTSD and CPTSD whereby PTSD was approximated with DSM-5 measures. Although this has been done in prior work (e.g., Cloitre et al., 2014), it is a limitation that none of our measures were developed to specifically assess the ICD-11 criteria for each disorder. Fourth, a limitation of all LCAs is their dependence on researcher interpretation of fit statistics and the difficulty of assessing the power to detect classes. Regarding the former, we used multiple statistics to deduce the best-fitting model, but the determination of which fit statistics are superior obfuscates absolute certainty. In addition, given that we had 23 observable indicators, our sample may have been underpowered. For this reason, a larger sample is recommended for future research.

Finally, our results should be contextualized by our sample. Our sample differed from previous studies in several important ways: It was more racially and ethnically diverse, participants were non–treatment-seeking, most individuals had subthreshold PTSD, and the sample consisted of urban-dwelling young adults. The diversity and non–treatment-seeking nature of our sample were study strengths, but it will be important for future researchers to replicate these findings among a sample that includes more diversity regarding age, socioeconomic and marital statuses, and sex, given that our sample was 71% female, as well as higher levels of symptom severity to examine whether a four-class solution characterizes samples with more severe symptoms. Similarly, it will be important to repeat this work in a sample that includes more representation of non–trauma-exposed individuals given that trauma exposure is not a requirement for the BPD diagnosis. Finally, the cross-sectional nature of the parent study limits the ability to examine the temporal stability of the classes. The constructs in the present study may have overlapped to a greater extent in this sample given that young adults are still developing and may show changes in symptom presentation with age. Analyses that use longitudinal data could reveal that individuals transition between classes.

Per the DSM-5, PTSD can be diagnosed in 636,120 different ways (Galatzer-Levy & Bryant, 2013), demonstrating the heterogeneity within this disorder and the need to examine whether additional diagnoses, such as CPTSD, overlap with the traditional diagnosis. The present study aimed to ascertain whether PTSD, CPTSD, and BPD could be distinguished by separate classes within a sample of racially and ethnically diverse young adults. Although prior LCAs have demonstrated successful separation, the present results showed contrary findings. In both the three- and four-class models, CPTSD and BPD failed to separate into distinct groups, highlighting the overlap between these two constructs. Accordingly, the present results highlight the limitations in categorizing CPTSD and BPD as distinct entities (e.g., Knefel et al., 2016) even though there is a clinical utility to a CPTSD diagnosis (Herman et al., 2012; Hyland et al., 2019). Notably, the present study depended on two diagnostic classification systems—the DSM-5 and the ICD-11—both which interpret diagnostic categories as discrete entities rather than diagnoses that may have overlaps in their boundaries or lie on a continuum (Cramer et al., 2010; Kotov et al., 2017). The mixed findings on the diagnostic separation of PTSD, CPTSD, and BPD in the literature and in the present study suggest that it may be more important to shift from examining disparate diagnoses (i.e., the disease model) to examining families of disorders with underlying qualities (e.g., HiTOP model; Kotov et al., 2017) or examining differences in the severity of presentations following trauma exposure (e.g., the Research Domain Criteria model; Clark et al., 2017). One could speculate whether CPTSD and BPD in the context of PTSD are more indicative of the severity of psychological distress after trauma exposure rather than a disparate constellation of symptoms (e.g., Hyland et al., 2019). Not only could this clarify diagnostic understanding but also lead to a systematic change in research and clinical practice whereby teams specializing in one disorder can work together rather than separately in understanding and treating families of psychopathologies. Research, such as the present study, that indicates ways in which these seemingly distinct constructs overlap and diverge will be important in guiding the field’s understanding of how to group and conceptualize families of psychopathology. Given that many studies have used similar methodologies as that used in the present study and found divergent results, a large-scale study with a sample that is diverse in age, race and ethnicity, and gender, is needed to yield substantial conclusions. Until then, these results add to the literature and suggest that CPTSD and BPD overlap in a young-adult sample and thus, may be capturing similar psychological profiles following trauma exposure.

Supplementary Material

Supplementary Materials

Open Practices Statement.

Neither study reported on in this article was formally preregistered. Neither the data nor the materials have been made available on a permanent third-party archive; requests for the data or materials should be sent via email to the lead author at saraiya@musc.edu

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing. [Google Scholar]
  2. Anda RF, Butchart A, Felitti VJ, & Brown DW (2010). Building a framework for global surveillance of the public health implications of adverse childhood experiences. American Journal of Preventive Medicine, 39(1), 93–98. 10.1016/j.amepre.2010.03.015 [DOI] [PubMed] [Google Scholar]
  3. Böttche M, Ehring T, Krüger-Gottschalk A, Rau H, Schäfer I, Schellong J, Dyer A, & Knaevelsrud C (2018). Testing the ICD-11 proposal for complex PTSD in trauma-exposed adults: Factor structure and symptom profiles. European Journal of Psychotraumatology, 9(1), 1512264. 10.1080/20008198.2018.1512264 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brewin CR, Cloitre M, Hyland P, Shevlin M, Maercker A, Bryant RA, Humayun A, Jones LM, Kagee A, Rousseau C, Somasundaram D, Suzuki Y, Wessely S, van Ommeren M, & Reed GM (2017). A review of current evidence regarding the ICD-11 proposals for diagnosing PTSD and complex PTSD. Clinical Psychology Review, 58, 1–15. 10.1016/j.cpr.2017.09.001 [DOI] [PubMed] [Google Scholar]
  5. Briere J, & Spinazzola J (2005). Phenomenology and psychological assessment of complex posttraumatic states. Journal of Traumatic Stress, 18(5), 401–412. [DOI] [PubMed] [Google Scholar]
  6. Candea D-M, & Szentagotai-Tatar A (2018). Shame-proneness, guilt-proneness and anxiety symptoms: A meta-analysis. Journal of Anxiety Disorders, 58, 78–106. 10.1016/j.janxdis.2018.07.005 [DOI] [PubMed] [Google Scholar]
  7. Clark LA, Cuthbert B, Lewis-Fernández R, Narrow WE, & Reed GM (2017). Three approaches to understanding and classifying mental disorder: ICD-11, DSM-5, and the National Institute of Mental Health’s Research Domain Criteria (RDoC). Psychological Science in the Public Interest, 18(2), 72–145. 10.1177/1529100617727266 [DOI] [PubMed] [Google Scholar]
  8. Cloitre M, Garvert DW, Brewin CR, Bryant RA, & Maercker A (2013). Evidence for proposed ICD-11 PTSD and complex PTSD: A latent profile analysis. European Journal of Psychotraumatology, 4(1), 20706. 10.3402/ejpt.v4i0.20706 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cloitre M, Garvert DW, Weiss B, Carlson EB, & Bryant RA (2014). Distinguishing PTSD, complex PTSD, and borderline personality disorder: A latent class analysis. European Journal of Psychotraumatology, 5(1), 25097. 10.3402/ejpt.v5.25097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cloitre M, Hyland P, Bisson JI, Brewin CR, Roberts NP, Karatzias T, & Shevlin M (2019). ICD-11 Posttraumatic stress disorder and complex posttraumatic stress disorder in the United States: A population-based study. Journal of Traumatic Stress, 32(6), 833–842. 10.1002/jts.22454 [DOI] [PubMed] [Google Scholar]
  11. Cloitre M, & Schmidt JA (2015). STAIR narrative therapy. In Schnyder U & Cloitre M (Eds.), Evidence based treatments for trauma-related psychological disorders: A practical guide for clinicians (p. 277–297). Springer International Publishing. 10.1007/978-3-319-07109-1_14 [DOI] [Google Scholar]
  12. Cloitre M, Stolbach BC, Herman JL, van der Kolk B, Pynoos R, Wang J, & Petkova E (2009). A developmental approach to complex PTSD: Childhood and adult cumulative trauma as predictors of symptom complexity. Journal of Traumatic Stress, 22(5), 399–408. 10.1002/jts.20444 [DOI] [PubMed] [Google Scholar]
  13. Cohen P, Chen H, Gordon K, Johnson J, Brook J, & Kasen S (2008). Socioeconomic background and the developmental course of schizotypal and borderline personality disorder symptoms. Development and Psychopathology, 20(2), 633–650. 10.1017/S095457940800031X [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cramer AOJ, Waldorp LJ, van der Maas HLJ, & Borsboom D (2010). Comorbidity: A network perspective. Behavioral and Brain Sciences, 33(2–3), 137–150. 10.1017/S0140525X09991567 [DOI] [PubMed] [Google Scholar]
  15. Derogatis LR (2001). Brief Symptom Inventory (BSI)-18. Administration, scoring, and procedures manual NCS Pearson, Inc. [Google Scholar]
  16. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, Koss MP, & Marks JS (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14(4), 245–258. 10.1016/s0749-3797(98)00017-8 [DOI] [PubMed] [Google Scholar]
  17. First MB, & Gibbon M (2004). The Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) and the Structured Clinical Interview for DSM-IV Axis II Disorders (SCID-II). Comprehensive handbook of psychological assessment, Vol. 2: Personality assessment (pp. 134–143). Wiley. [Google Scholar]
  18. Ford J (2015). An affective cognitive neuroscience-based approach to PTSD psychotherapy: The TARGET model. Journal of Cognitive Psychotherapy, 29(1), 68–91. 10.1891/0889-8391.29.1.68 [DOI] [PubMed] [Google Scholar]
  19. Ford JD, & Courtois CA (2009). Defining and understanding complex trauma and complex traumatic stress disorders. In Courtois CA & Ford JD (Eds.), Treating complex traumatic stress disorders: An evidence-based guide (pp. 13–30). Guilford Press. [Google Scholar]
  20. Frost R, Hyland P, Shevlin M, & Murphy J (2018). Distinguishing complex PTSD from borderline personality disorder among individuals with a history of sexual trauma: A latent class analysis. European Journal of Trauma & Dissociation, S2468749918300255. 10.1016/j.ejtd.2018.08.004 [DOI]
  21. Galatzer-Levy I, & Bryant R (2013). 636,120 ways to have posttraumatic stress disorder. Perspectives on Psychological Science, 8(6), 651–662. 10.1177/1745691613504115 [DOI] [PubMed] [Google Scholar]
  22. Gillespie CF, Bradley B, Mercer K, Smith AK, Conneely K, Gapen M, Weiss T, Schwartz AC, Cubells JF, & Ressler KJ (2009). Trauma exposure and stress-related disorders in inner-city primary care patients. General Hospital Psychiatry, 31(6), 505–514. 10.1016/j.genhosppsych.2009.05.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gray MJ, Litz BT, Hsu JL, & Lombardo TW (2004). Psychometric properties of the Life Events Checklist. Assessment, 11(4), 330–341. 10.1177/1073191104269954 [DOI] [PubMed] [Google Scholar]
  24. Harned MS, & Linehan MM (2008). Integrating dialectical behavior therapy and prolonged exposure to treat co-occurring borderline personality disorder and PTSD: Two case studies. Cognitive and Behavioral Practice, 15(3), 263–276. 10.1016/j.cbpra.2007.08.006 [DOI] [Google Scholar]
  25. Herman J (1992). Trauma and recovery: The aftermath of violence—From domestic violence to political terror Guilford Press. [Google Scholar]
  26. Hyland P, Shevlin M, Brewin CR, Cloitre M, Downes AJ, Jumbe S, Karatzias T, Bisson JI, & Roberts NP (2017). Validation of posttraumatic stress disorder (PTSD) and complex PTSD using the International Trauma Questionnaire. Acta Psychiatrica Scandinavica, 136(3), 313–322. 10.1111/acps.12771 [DOI] [PubMed] [Google Scholar]
  27. Hyland P, Karatzias T, Shevlin M, & Cloitre M (2019). Examining the discriminant validity of complex posttraumatic stress disorder and borderline personality disorder symptoms: Results from a United Kingdom population sample. Journal of Traumatic Stress, 32(6), 855–863. 10.1002/jts.22444 [DOI] [PubMed] [Google Scholar]
  28. Jacobsberg L, Perry S, & Frances A (1995). Diagnostic agreement between the SCID-II screening questionnaire and the Personality Disorder Examination. Journal of Personality Assessment, 65(3), 428–433. 10.1207/s15327752jpa6503_4 [DOI] [PubMed] [Google Scholar]
  29. Jowett S, Karatzias T, Shevlin M, & Albert I (2019). Differentiating symptom profiles of ICD-11 PTSD, complex PTSD, and borderline personality disorder: A latent class analysis in a multiply traumatized sample. Personality Disorders, 11(1), 36–45. 10.1037/per0000346 [DOI] [PubMed] [Google Scholar]
  30. Karatzias T, Hyland P, Bradley A, Fyvie C, Logan K, Easton P, Thomas J, Philips S, Bisson JI, Roberts NP, Cloitre M, & Shevlin M (2019). Is self-compassion a worthwhile therapeutic target for ICD-11 Complex PTSD (CPTSD)? Behavioural and Cognitive Psychotherapy, 47(3), 257–269. 10.1017/S1352465818000577 [DOI] [PubMed] [Google Scholar]
  31. Knefel M, Garvert DW, Cloitre M, & Lueger-Schuster B (2015). Update to an evaluation of ICD-11 PTSD and complex PTSD criteria in a sample of adult survivors of childhood institutional abuse by Knefel & Lueger-Schuster (2013): A latent profile analysis. European Journal of Psychotraumatology, 6(1), 25290. 10.3402/ejpt.v6.25290 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Knefel M, Tran US, & Lueger-Schuster B (2016). The association of posttraumatic stress disorder, complex posttraumatic stress disorder, and borderline personality disorder from a network analytical perspective. Journal of Anxiety Disorders, 43, 70–78. 10.1016/j.janxdis.2016.09.002 [DOI] [PubMed] [Google Scholar]
  33. Kotov R, Krueger RF, Watson D, Achenbach TM, Althoff RR, Bagby RM, Brown TA, Carpenter WT, Caspi A, Clark LA, Eaton NR, Forbes MK, Forbush KT, Goldberg D, Hasin D, Hyman SE, Ivanova MY, Lynam DR, Markon K, … Zimmerman M (2017). The Hierarchical Taxonomy of Psychopathology (HiTOP): A dimensional alternative to traditional nosologies. Journal of Abnormal Psychology, 126(4), 454–477. 10.1037/abn0000258 [DOI] [PubMed] [Google Scholar]
  34. Landy MSH, Wagner AC, Brown-Bowers A, & Monson CM (2015). Examining the evidence for complex posttraumatic stress disorder as a clinical diagnosis. Journal of Aggression, Maltreatment & Trauma, 24(3), 215–236. 10.1080/10926771.2015.1002649 [DOI] [Google Scholar]
  35. Linehan MM (1993). Cognitive-behavioral treatment of borderline personality disorder Guilford Press. [Google Scholar]
  36. López-Castro T, Saraiya T, Zumberg-Smith K, & Dambreville N (2019). Association between shame and posttraumatic stress disorder: A meta-analysis. Journal of Traumatic Stress, 32(4), 484–495. 10.1002/jts.22411 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Luyten P, Fontaine JR, Corveleyn J (2002). Does the Test of Self-Conscious Affect (TOSCA) measure maladaptive aspects of guilt and adaptive aspects of shame? An empirical investigation. Personality and Individual Differences, 33 (8), 1373–1387. 10.1016/S0191-8869(02)00197-6 [DOI] [Google Scholar]
  38. McBride O, Hyland P, Murphy J, & Elklit A (2019). Network analysis of posttraumatic stress experiences of adults seeking psychological treatment for childhood sexual abuse. Journal of Traumatic Stress, 33(1), 10–18. 10.1002/jts.22381 [DOI] [PubMed] [Google Scholar]
  39. McDonald SD, & Calhoun PS (2010). The diagnostic accuracy of the PTSD Checklist: A critical review. Clinical Psychology Review, 30(8), 976–987. 10.1016/j.cpr.2010.06.012 [DOI] [PubMed] [Google Scholar]
  40. Monson CM, Fredman SJ, & Dekel R (2010). Posttraumatic stress disorder in an interpersonal context. In Beck JG (Ed.), Interpersonal processes in the anxiety disorders: Implications for understanding psychopathology and treatment (pp. 179–208). American Psychiatric Association. 10.1037/12084-007 [DOI] [Google Scholar]
  41. Muthén LK, & Muthén BO (1998). Mplus user’s guide (8th ed.). Muthén & Muthén. [Google Scholar]
  42. Perkonigg A, Hofler M, Cloitre M, Wittchen HU, Trautmann S, & Maercker A (2016). Evidence for two different ICD-11 posttraumatic stress disorders in a community sample of adolescents and young adults. European Archives of Psychiatry and Clinical Neuroscience, 266(4), 317–328. 10.1007/s00406-015-0639-4 [DOI] [PubMed] [Google Scholar]
  43. Peters JR & Geiger PJ (2016). Borderline personality disorder and self-conscious affect: Too much shame but not enough guilt? Personality Disorders, 7(3), 303–308. 10.1037/per0000176 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Powers A, Fani N, Carter S, Cross D, Cloitre M, & Bradley B (2017). Differential predictors of DSM-5 PTSD and ICD-11 complex PTSD among African American women. European Journal of Psychotraumatology, 8(1), 1338914. 10.1080/20008198.2017.1338914 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pugh LR, Taylor PJ, & Berry K (2015). The role of guilt in the development of post-traumatic stress disorder: A systematic review. Journal of affective disorders, 182, 138–150. 10.1016/j.jad.2015.04.026 [DOI] [PubMed] [Google Scholar]
  46. Resick PA, Williams LF, Suvak MK, Monson CM, & Gradus JL (2012). Long-term outcomes of cognitive–behavioral treatments for posttraumatic stress disorder among female rape survivors. Journal of Consulting and Clinical Psychology, 80(2), 201–210. 10.1037/a0026602 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Saraiya T, Fareri D, López-Castro T, Hien D, Fertuck E, & Melara R (2019). The social cognitive appraisal of trustworthiness in individuals with dimensional levels of post-traumatic stress symptoms: a translational study. European Journal of Psychotraumatology, 10(1), 1697582. 10.1080/20008198.2019.1697582 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sodano SM, & Tracey T (2011). A brief inventory of interpersonal problems-circumplex using nonparametric item response theory: Introducing the IIP-C-IRT. Journal of Personality Assessment, 93(1), 62–75. 10.1080/00223891.2010.528482 [DOI] [PubMed] [Google Scholar]
  49. Spatz Widom C, Czaja SJ, & Paris J (2009). A prospective investigation of borderline personality disorder in abused and neglected children followed up into adulthood. Journal of Personality Disorders, 23(5), 433–446. 10.1521/pedi.2009.23.5.433 [DOI] [PubMed] [Google Scholar]
  50. Tangney JP, Wagner PE, & Gramzow R (1989). The Test of Self-Conscious Affect (TOSCA) George Mason University. [Google Scholar]
  51. Tangney JP (1995). Recent advances in the empirical study of shame and guilt. American Behavioral Scientist, 38(8), 1132–1145. 10.1177/0002764295038008008 [DOI] [Google Scholar]
  52. Turniansky H, Ben-Dor D, Krivoy A, Weizman A, & Shoval G (2019). A history of prolonged childhood sexual abuse is associated with more severe clinical presentation of borderline personality disorder in adolescent female inpatients: A naturalistic study. Child Abuse & Neglect, 98, 104222. 10.1016/j.chiabu.2019.104222 [DOI] [PubMed] [Google Scholar]
  53. Vater A, Schröder-Abé M, Weißgerber S, Roepke S, & Schütz A (2015). Self-concept structure and borderline personality disorder: Evidence for negative compartmentalization. Journal of Behavior Therapy and Experimental Psychiatry, 46, 50–58. 10.1016/j.jbtep.2014.08.003 [DOI] [PubMed] [Google Scholar]
  54. Weathers FW, Blake DD, Schnurr PP, Kaloupek DG, Marx BP, & Keane TM (2013). The Life Events Checklist for DSM-5 (LEC-5) https://www.ptsd.va.gov/professional/assessment/te-measures/life_events_checklist.asp
  55. Weathers FW, Litz BT, Keane TM, Palmieri PA, Marx BP, & Schnurr PP (2013). The PTSD Checklist for DSM-5 (PCL-5) https://www.ptsd.va.gov/professional/assessment/adult-sr/ptsd-checklist.asp
  56. Wiesner M, Chen V, Windle M, Elliott MN, Grunbaum JA, Kanouse DE, & Schuster MA (2010). Factor structure and psychometric properties of the Brief Symptom Inventory-18 in women: A MACS approach to testing for invariance across racial/ethnic groups. Psychological Assessment, 22 (4), 912–922. 10.1037/a0020704 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wolf EJ, Miller MW, Kilpatrick D, Resnick HS, Badour CL, Marx BP, Keane TM, Rosen RC, & Friedman MJ (2014). ICD-11 complex PTSD in U.S. national and veteran samples: Prevalence and structural associations with PTSD. Clinical Psychological Science, 3(2), 215–229. 10.1177/2167702614545480 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. World Health Organization. (2018). International classification of diseases for mortality and morbidity statistics (11th rev.) https://icd.who.int/browse11/l-m/en
  59. Zerach G, Shevlin M, Cloitre M, & Solomon Z (2019). Complex posttraumatic stress disorder (CPTSD) following captivity: A 24-year longitudinal study. European Journal of Psychotraumatology, 10(1), 1616488. 10.1080/20008198.2019.1616488 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Materials
NIHMS1855158-supplement-Supplementary_Materials.docx (23.9KB, docx)

RESOURCES