Subthreshold post-traumatic stress disorder (PTSD), which is not a recognised diagnostic entity, is highly prevalent in trauma-exposed civilian and military populations worldwide, often two to three times more prevalent than PTSD. For example, a nationally representative study of USA veterans found that 22.1% met criteria for subthreshold PTSD in their lifetimes and 13.5% in the past month, which was substantially higher than the 8% and 4.5% prevalence of lifetime and past-month PTSD, respectively.Reference Mota, Tsai and Sareen1 In addition to being prevalent, subthreshold PTSD is associated with compromised mental and physical health, as well as functional impairment.Reference Pietrzak, Goldstein, Southwick and Grant2 A recent meta-analysis found that, relative to trauma-exposed individuals without PTSD symptoms, those with subthreshold PTSD had elevated rates of psychiatric comorbidities, such as depressive symptoms, suicidality and substance abuse; reduced social and occupational functioning; and greater utilisation of healthcare services.Reference Brancu, Mann-Wrobel and Beckham3 Subthreshold PTSD is also linked to higher rates of suicidal ideation above and beyond the effect of major depressive disorder.Reference Marshall, Olfson and Hellman4
Despite the robust association of subthreshold PTSD with psychiatric comorbidities, suicidality and functional impairment, it has generally been ignored both in clinical settings and in the context of compensation.Reference Myelle and Maes5 To date, it remains unknown whether subthreshold PTSD may represent a risk factor for the development of PTSD in representative samples, and which factors may predict the conversion from subthreshold PTSD to PTSD. Given the chronicity, impairment and high cost associated with treating PTSD, examining whether subthreshold PTSD is a potential risk factor for PTSD could help inform prognostic models of PTSD and population-based prevention efforts to mitigate risk of developing this disorder.
Method
Sample
Data were analysed from the National Health and Resilience in Veterans Study (NHRVS), a 7-year, nationally representative, prospective cohort study of 3157 USA veterans. The sample was obtained from KnowledgePanel, a survey research panel representing approximately 98% of USA households that is maintained by GfK Custom Research (now Ipsos). A baseline (i.e. wave 1) survey was conducted in 2011 and follow-up surveys were conducted in 2013, 2015 and 2018. A total of 2305 (73%) veterans completed at least one follow-up assessment (mean number of follow-up assessments 2.2, s.d. 0.8, range 1–3 assessments). The Trauma History ScreenReference Carlson, Smith and Palmieri6 and PTSD Checklist–Specific Stressor Version (PCL-S)Reference Weathers, Litz, Herman, Huska and Keane7 were administered to assess trauma history and PTSD symptoms (see Table 1). Several sociodemographic, military, health and psychosocial variables associated with PTSDReference Wisco, Marx and Wolf8,Reference Mota, Cook and Smith9 were assessed as potential determinants of PTSD in veterans with subthreshold PTSD at wave 1, over the 7-year study period (Table 1). Wave 1 and follow-up PCL-S data were available for a total of 2155 veterans. Post-stratification weights were applied in inferential analyses, to permit generalisability of results to the USA veteran population.
Demographic, trauma and clinical characteristics by post-traumatic stress disorder (PTSD) status over the 7-year follow-up period
The Trauma History Screen was used to assess exposures to 13 potentially traumatic life event types, with the option to additional endorse an ‘other’ event.Reference Carlson, Smith and Palmieri6 An adapted lifetime version of the PTSD Checklist–Specific Stressor Version was used to assess lifetime PTSD (criteria A–F).Reference Weathers, Litz, Herman, Huska and Keane7 Adapted modules from the Mini Neuropsychiatric Interview were used to assess lifetime major depressive, alcohol and drug use disorders.Reference Sheehan, Lecrubier and Sheehan11 The Three-Item Loneliness Scale was used to assess loneliness; scores on each item range from 1 (hardly ever) to 3 (often), and are summed to yield a total score ranging from 3–9, with higher scores reflecting greater loneliness.Reference Hughes, Waite, Hawkley and Cacioppo12 The Medical Outcomes Study Cognitive Functioning Scale-Revised was used to assess cognitive functioning; scores range from 0 to 100, with lower scores indicating worse cognitive functioning.Reference Stewart, Ware, Sherbourne, Wells, Stewart and Ware13 A Medical Conditions Checklist derived from the Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV14 was used to assess the presence/absence of 20 healthcare professional-diagnosed medical conditions (e.g. diabetes, heart disease, migraine); a higher number of conditions indicates greater medical burden. The following item from the Life Orientation Test-Revised was used to assess dispositional optimism (‘In uncertain times, I usually expect the best’); scores ranges from 1 (strongly disagree) to 7 (strongly agree).Reference Scheier, Carver and Bridges15 The Purpose in Life Test-Short Form was used to assess purpose in life; scores range from 4 to 28, with higher scores indicating greater purpose in life.Reference Schulenberg, Schnetzer and Buchanan16 An abbreviated five-item version of the Medical Outcome Study Social Support Scale was used to assess perceived social support; scores range from 5 to 25, with higher scores indicating greater perceived social support.Reference Sherbourne and Stewart17
a. Row percentages are shown for wave 1 PTSD status; percentages for all other variables are column percentages.
Ethics statement
The NHRVS was approved by the Human Subjects Subcommittee of the VA Connecticut Healthcare System (protocol number RP0002). All participants provided informed consent.
Assessments
The DSM-IV version of the PCL-S was used to assess PTSD symptoms related to veterans’ ‘worst’ traumatic event as assessed by the Trauma History Screen at wave 1. The most prevalent worst traumatic events among veterans with subthreshold PTSD were life-threatening illness or injury (23.5%), sudden death of a close family member or friend (21.8%) and military-related trauma (8.8%). PTSD symptom endorsement was operationalized as reporting being bothered ‘moderately’, ‘quite a bit’ or ‘extremely’, by the symptom. Subthreshold PTSD was classified as meeting symptom criteria for cluster B (re-experiencing) plus cluster C (avoidance) or D (hyperarousal), or those who met criteria for cluster B plus at least one cluster C and one cluster D symptom.Reference Mota, Tsai and Sareen1,Reference Brancu, Mann-Wrobel and Beckham3 In waves 2–4, the DSM-5 version of the PCL (PCL-5)Reference Weathers, Litz, Keane, Palmieri, Marx and Schnurr10 was used to assess PTSD symptoms related to veterans’ worst traumatic event endorsed at baseline. Incident PTSD was defined as meeting DSM-5 criteria A (stressor), B (intrusion symptoms), C (avoidance), D (negative mood and cognitions), E (hyperarousal), F (duration; additional question added to the PCL-5: ‘How long did these reactions last?’, with criterion F met if symptoms lasted >1 month) and G (clinically significant distress/functional impairment; additional question added to the PCL-5: ‘Did these reactions cause you distress or result in a failure to fulfill obligations at home, work, or school?’ Endorsement of ‘moderately’, ‘quite a bit’ or ‘extremely’ was indicative of meeting criterion G).
Data analysis
Chi-squared analyses and independent samples t-tests were conducted to compare demographic, trauma and clinical characteristics of veterans with and without PTSD over the 7-year follow-up period. Multivariable binary logistic regression analyses were then conducted in veterans with subthreshold PTSD at wave 1, to evaluate specific PTSD symptoms associated with incident PTSD (backward Wald estimation method); and sociodemographic, military, health and psychosocial determinants of incident PTSD (see Table 1 for variables included in this analysis). Relative importance analyses,Reference Tonidandel and LeBreton18 which partition explained variance among independent variables while accounting for intercorrelations among them, were then conducted to determine the relative variance in incident PTSD that was explained by significant predictors in these analyses.
Results
Table 1 shows demographic, trauma and clinical characteristics of the sample by PTSD status over the 7-year follow-up period. In the full sample, a total of 214 veterans (weighted 10.9%) screened positive for full PTSD at one or more follow-up assessments over the 7-year follow-up period. As shown in Table 1, of the veterans with subthreshold PTSD at wave 1 (n = 112, 5.3% of the wave 1 cohort), 34.3% developed PTSD at one or more assessments, relative to 7.6% of trauma-exposed veterans without subthreshold PTSD (n = 1983, 90.6% of the wave 1 cohort, relative risk ratio 6.4, 95% CI 4.1–9.9, P < 0.001); of veterans with full PTSD at wave 1 (n = 60, 4.1% of the wave 1 cohort), 54.3% screened positive for full PTSD at one or more follow-up assessments.
In analyses excluding veterans with lifetime PTSD (n = 127, 6.4% of the wave 1 cohort), veterans with subthreshold PTSD at wave 1 had comparably elevated risk of incident PTSD over the follow-up period (34.1% v. 6.3%; relative risk ratio 7.6, 95% CI 4.7–12.3); and in an analysis adjusted for wave 1 PTSD, major depressive disorder and alcohol and drug use disorder, wave 1 subthreshold PTSD was associated with a five-fold greater likelihood of developing PTSD over the follow-up period (relative risk ratio 5.0, 95% CI 3.1–8.1).
Among veterans with subthreshold PTSD at wave 1, positive endorsement of three PTSD symptoms emerged as independent predictors of incident PTSD: psychogenic amnesia (i.e. ‘trouble remembering important parts of trauma’; relative risk ratio 5.44, 95% CI 1.43–20.67), hypervigilance (i.e. ‘being super alert or watchful or on guard’; relative risk ratio 3.98, 95% CI 1.52–10.40) and trauma-related nightmares (i.e. ‘repeated, disturbing dreams of trauma’; relative risk ratio 2.98, 95% CI 1.07–8.26). The total model R 2 was 0.31; hypervigilance (41.8% relative variance explained) explained most of the variance in incident PTSD, and psychogenic amnesia (32.2% relative variance explained) and trauma-related nightmares (26% relative variance explained) explained the remainder of the variance in this outcome
Among veterans with subthreshold PTSD at wave 1, greater age (relative risk ratio 1.05, 95% CI 1.01–1.10), number of traumas since wave 1 (relative risk ratio 1.22, 95% CI 1.11–2.11) and cognitive difficulties (relative risk ratio 1.03, 95% CI 1.01–1.06) at wave 1 were independently associated with increased risk for developing PTSD, whereas greater dispositional optimism at wave 1 was associated with reduced risk of developing PTSD (relative risk ratio 0.61, 95% CI 0.43–0.88). The total model R 2 was 0.37; greater number of traumas since wave 1 (65.5% relative variance explained) explained the majority of variance in incident PTSD, whereas cognitive difficulties (18% relative variance explained), dispositional optimism (12.2% relative variance explained) and age (4.3% relative variance explained) explained the remainder of the variance in this outcome.
Discussion
Using data from a nationally representative sample of USA veterans, results of this study provide evidence for the potential prognostic utility of subthreshold PTSD, and identify targets for population-based preventive interventions for PTSD in this population. Results revealed that subthreshold PTSD, which was more prevalent than full PTSD at baseline (5.3% v. 4.1%), was associated with a more than six-fold greater likelihood of developing PTSD over a 7-year follow-up period. Further, among veterans with subthreshold PTSD, hypervigilance, psychogenic amnesia and trauma-related nightmares were associated with significantly increased risk for the development of full PTSD. This finding suggests that greater sensitisation to trauma and possibly the emergence/exacerbation of PTSD symptoms in late lifeReference Mota, Tsai and Kirwin19 may, in part, drive the development of PTSD in veterans with subthreshold PTSD. It further underscores the importance of assessing individual PTSD symptoms across symptom clusters in predicting risk for this disorder in this population.
Among veterans with subthreshold PTSD at wave 1, greater age, cognitive difficulties and lower dispositional optimism at wave 1, and increased trauma burden over the follow-up period, were associated with increased risk of developing PTSD. Increased trauma burden over the follow-up period explained nearly two-thirds of the variance in incident PTSD, which suggests that subthreshold PTSD may sensitise veterans to the deleterious effects of new-onset traumatic life events, which in turn increases risk of developing full PTSD; alternatively, new-onset traumas may give rise to PTSD in relation to a new traumatic event. Collectively, these results suggest that efforts to assess, monitor and treat negative psychological effects of trauma, mitigate cognitive difficulties and promote dispositional optimism20-22 may help mitigate risk of developing PTSD in veterans with subthreshold PTSD.
Limitations of this study include the assessment of measures via a web-based platform that relied on self-report data; and its focus on a demographically homogeneous sample of USA veterans. Further research, using clinician-administered diagnostic interviews in more diverse samples, is needed to evaluate the generalisability of the results reported herein.
Notwithstanding these limitations, results of this study suggest that subthreshold PTSD may represent a risk factor for PTSD, as it was associated with a more than six-fold greater likelihood of developing PTSD. They further underscore the importance of pursuing replication of these results in other samples, using clinical interviews and DSM-5- and ICD-11-based definitions of subthreshold PTSD; raising awareness of subthreshold PTSD and related syndromes, such as adjustment disorder, as potential risk factors for PTSD; and evaluating the efficacy of targeted preventive interventions to help reduce risk for the more chronic, difficult-to-treat and costly manifestation of full PTSD in veterans and other trauma-affected populations.
Data availability
The data that support the findings of this study are available upon reasonable request from the corresponding author, R.H.P. The data are not publicly available due to their containing information that could compromise the privacy of research participants.
Acknowledgements
The authors thank the veterans who participated in the National Health and Resilience in Veterans Study and the Ipsos staff who facilitated data collection, particularly Sergei Rodkin, PhD, Robert Torongo, MA and Alyssa Marciniak, MA.
Author contributions
R.H.P., J.H.K. and S.M.S. designed the National Health and Resilience in Veterans Study and acquired funding. R.H.P. conducted the data analyses. R.H.P., F.J.G. and S.M.S. conceptualised the study and drafted the initial version of the manuscript. All authors interpreted the results, provided critical revisions of the manuscript and approved the final version.
Funding
The National Health and Resilience in Veterans Study is supported by the U.S. Department of Veterans Affairs National Center for Posttraumatic Stress Disorder. Data collection for the 7-year follow-up survey was supported in part by National Institute on Aging grant U01AG032284 awarded to Becca Levy, PhD as principal investigator and R.H.P. as co-investigator.
Declaration of interest
None.
eLetters
No eLetters have been published for this article.