Blue lines represent positive edges and red lines represent negative edges. The width and hue of the lines indicate relative edge strength. Aa indicates anxious arousal; An, anhedonia; Av, avoidance; Da, dysphoric arousal; Eb, externalizing behavior; and Na, negative affect.
eTable. Mappings of PTSD Checklist for DSM-5 Items for DSM-5, 7-Factor, and 8-Factor Structural Models
Customize your JAMA Network experience by selecting one or more topics from the list below.
Duek O, Spiller TR, Rubenstein A, Pietrzak RH, Harpaz-Rotem I. Exploration of a Novel Model of Intrusive Symptoms in Posttraumatic Stress Disorder Among US Veterans. JAMA Netw Open. 2022;5(3):e223555. doi:10.1001/jamanetworkopen.2022.3555
Posttraumatic stress disorder (PTSD) is a heterogeneous disorder that has been plagued by a lack of clarity regarding its core symptom dimensions. While intrusions represent the core, defining symptom cluster of PTSD, 2 potentially distinct processes have recently been proposed to comprise these symptoms: internally generated symptoms (eg, flashbacks and nightmares); and externally generated symptoms (eg, physiological/emotional reactivity following trauma reminders).1
However, to our knowledge, no study has evaluated whether this more refined phenotypic model of intrusive symptoms provides a better empirical fit to PTSD symptom data than current models, such as the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) (DSM-5)2 4-cluster model and the 7-factor model, which proposes hybrid anhedonia and hypervigilance clusters.3
To address this gap, we evaluated a PTSD model that divides intrusions into internally and externally cued symptoms. We compared the fit of this model with the DSM-5 and 7-factor models. We then examined how these 2 intrusion clusters correlated to other PTSD symptom clusters.
This cross-sectional study was approved by the West Haven VA institutional review board with a waiver of informed consent because we used deidentified data. This study reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Two independent samples were analyzed. The first sample included veterans who completed the PTSD Checklist for DSM-5 (PCL-5)4 at admission into intensive PTSD treatment at the Department of Veterans Affairs (VA) between October 2014 and September 2016. The second sample (validation) included PCL-5 records of veterans who completed the PCL-5 on entering PTSD treatment at the VA between October 2018 and September 2019. The PCL-5 assesses the 20 DSM-5 PTSD symptoms. Cronbach α was 0.90 for sample 1 and 0.93 for sample 2.
We used latent network modeling, a framework in which latent variables are represented by nodes, and edges between them represent partial correlations (adjusted for the influence of all other nodes).5 We compared 3 different structural models of PTSD symptoms: (1) DSM-5, (2) 7-factor model, and (3) 8-factor (internally and externally cued intrusions) (eTable in the Supplement). Model fit was independently tested and compared in samples 1 and 2 (validation). All analyses were conducted in R statistical software with the Psychonetrics package version 0.9 (R Project for Statistical Computing). P values were 2-sided, and significance was set a α = .10. Data were analyzed from August to October 2021.
Sample 1 included 7367 veterans (median [IQR] age, 44.1 [33.3-55.1] years, 6255 [85%] men). Sample 2 included 52 609 PCL-5 records (median [IQR] age, 44.7 [35.3-56.8] years; 45 193 [86%] men). The mean (SD) PCL-5 score was 59.1 (11.9) in sample 1 and 57.8 (12.5) in sample 2. The proposed 8-factor model (separating the intrusion cluster into internal and external) had significantly superior fit than the DSM-5 and 7-factor models (χ2 difference: sample one, 3811; sample two, 822; P < .001) (Table). Fit testing in the validation sample yielded comparable results, with the 8-factor model evidencing the best fit (Table). Network analysis revealed positive partial correlations between externally cued intrusions and avoidance symptoms and between internally cued intrusions and dysphoric arousal symptoms (Figure).
This cross-sectional study provides the first empirical support, to our knowledge, for a recently proposed conceptual distinction between internally and externally generated intrusive symptoms of PTSD.1 Latent network analyses revealed a superior fit of a theory-driven model with 2 separate intrusion clusters, as well as distinct connections between these 2 clusters, and other PTSD symptom clusters. Specifically, externally cued intrusions were associated with avoidance behaviors, while internally generated intrusions were associated with dysphoric arousal.
This distinction between intrusive symptoms may have clinical implications. For example, externally cued intrusions may have fear-based origins that respond better to exposure therapy, while internally generated intrusions may be linked to more cognitive-ruminative processes and thus, will need a different therapeutic approach.
While the cross-sectional and veteran-specific samples limit the generalizability of this study, results suggest that separating intrusive symptoms may provide a more nuanced phenotypic representation of the phenotypic expression of PTSD. Further research is needed to examine neurobiological underpinnings of these symptom clusters, and their utility in informing the prognosis of PTSD and precision medicine-based interventions for this disorder.
Accepted for Publication: February 2, 2022.
Published: March 21, 2022. doi:10.1001/jamanetworkopen.2022.3555
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Duek O et al. JAMA Network Open.
Corresponding Authors: Or Duek, PhD, Department of Psychiatry, Yale School of Medicine, 300 George St, New-Haven, CT (firstname.lastname@example.org); Ilan Harpaz-Rotem, PhD, ABPP, 200 Yale West Campus Dr, Orange, CT (email@example.com).
Author Contributions: Dr Harpaz-Rotem had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Duek, Harpaz-Rotem.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Duek, Rubenstein, Harpaz-Rotem.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Duek, Spiller.
Obtained funding: Harpaz-Rotem.
Administrative, technical, or material support: Rubenstein, Pietrzak.
Supervision: Pietrzak, Harpaz-Rotem.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Spiller was supported by an Early Postdoc Mobility Fellowship of the Swiss National Science Foundation (grant No. P2ZHP3_195191). Drs Pietrzak and Harpaz-Rotem were supported by grant No. CX-001538 from the VA Clinical Science Research and Development, and Dr Harpaz-Rotem was supported by grant No. MH115110 from the National Institute of Mental Health.
Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The views expressed in this article are solely those of the authors and do not reflect an endorsement by or the official policy or position of the US Department of Veterans Affairs or the US Government.