Most individuals exposed to a traumatic event do not develop post-traumatic stress disorder (PTSD); however, many individuals may experience varying amounts of sub-clinical levels of post-traumatic stress symptoms (PTSS). Notable differences in the presence and severity of PTSS have been found among individuals reporting traumatic events that would appear to be comparably stressful 123. This raises the possibility that some individuals may have pre-trauma vulnerabilities for developing PTSS or diagnosable PTSD.

Increased vulnerability to PTSS has been found to be associated with individual difference factors such as: personal history of depressive disorders 45, anxiety disorders 4 as well as introversion 6 and neuroticism 467. Underlying these various vulnerabilities may be a common factor termed, “negative affectivity.” Clark, Watson and Mineka 8 have described negative affectivity as a “…stable, heritable, and highly general trait dimension with a multiplicity of aspects ranging from mood to behavior” that is “…broadly related to psychopathology, including indices of both anxiety and depression…” (p. 104). Thus, negative affectivity would encompass a range of characteristics demonstrated to be highly correlated with each other and potentially with PTSS, including: anxiety, introversion, neuroticism, electrodermal lability, slowed habituation rate and increased conditionability 679.

Numerous studies have demonstrated concurrent associations between psychologic and psychophysiologic factors and PTSD symptoms. This research has primarily used cross-sectional study designs, i.e., comparisons are made between individuals with and without PTSD. From such studies it is difficult, if not impossible, to determine whether the observed differences are a consequence of developing PTSD or might represent pre-existing differences that are associated with risk for developing PTSD. We examined the PTSD literature to identify well-standardized psychologic and psychophysiologic measures that have been found to discriminate between individuals with and without PTSD in at least one or more published studies and that might serve as risk markers for the development of PTSD. The candidate variables we selected from this review as potential risk markers, in addition to age, education, gender and race, are listed in Table 1.

Note: SCL-90-R GSI = Symptom Checklist 90-Revised, Global Severity Score; BDI = Beck Depression Inventory; WAIS-R EST-IQ = Estimated WAIS-R Full Scale IQ from Shipley Institute of Living Scale; STAI-S = State/Trait Anxiety Inventory-State score; bpm = beats per minute; S = Siemens; V = volts; SC = skin conductance; HR = heart rate; Orienting Response = mean response to first presentation of the CS + and the CS- during the habituation phase; Acquisition and extinction response to CS + trials = mean response to CS + trials during the acquisition and extinction phases, respectively; habituation, acquisition, and extinction differential responses = the mean CS interval response to CS + trials minus the mean CS interval response to CS- trials for the habituation, acquisition, and extinction phases, respectively; CS = conditioned stimulus.

Briefly, studies of psychologic measures have found individuals with, compared to without, PTSD to have greater: general psychiatric symptoms e.g., 10111213, depression symptoms e.g., 1112141516, state and/or trait anxiety e.g., 1112131415171819 and neuroticism e.g., 1620212223. Whereas, greater: IQ 1624252627, extraversion 2228, openness 29 and agreeableness 2223 have been observed in individuals without, compared to with, PTSD. Of particular note, research cited above suggests that lower IQ and the personality traits of increased neuroticism, lower extraversion and lower agreeableness are pre-trauma predictors of increased risk for PTSD and/or PTSS.

Studies of psychophysiologic reactivity to startling stimuli and fear conditioning have reported several reliable differences between individuals with and without PTSD. The startle response represents a form of unconditioned reactivity to sudden, unexpected stimuli such as a loud noise that is mediated by brainstem neural circuitry 30 but is influenced by higher psychological processes such as emotional state. Magnitude of the eyeblink response is commonly used as the prototypical measure of startle in humans. However, startling acoustic stimuli can also produce increased skin conductance (SC) and heart rate (HR) responses, indicating that there are autonomic, as well as muscular, features of startle. The magnitude of reactivity to startling stimuli and rate of decline in response magnitude over repeated presentations of a startle stimulus, i.e., habituation, have been extensively used as measures of anxiety and/or emotional state. There is a substantial literature demonstrating larger magnitude: eyeblink e.g., 173132, skin conductance e.g., 193233 and heart rate e.g., 1017193133343536 responses to loud noise or tone stimuli in individuals with, compared to without, PTSD. Of particular note, findings from a recent prospective study of firefighter trainees suggests that eyeblink and SC response magnitudes to loud tones predict the development of greater PTSS following exposure to a traumatic event 37. In contrast, there is now compelling evidence that larger HR response magnitude to loud tones is an acquired feature of PTSD 3638. Slower habituation of SC responses e.g., 1017333536, HR responses 34 and eyeblink responses 38 have also been observed in individuals with PTSD. There is some evidence that slower habituation of SC and eyeblink responses, as measured by the number of trials to no-response, may be acquired features of PTSD 38.

Psychophysiologic studies of acquisition and extinction of conditioned fear responses have also demonstrated differences between individuals with and without PTSD. Fear conditioning provides a measure of an individual's propensity to associate a neutral cue with an aversive outcome. This propensity can be assessed from the strength of the conditioned response that is established during acquisition, as well as from its resistance to extinction. Measures of SC activity have been commonly used to assess human fear conditioning; measures of HR and facial EMG activity have been less frequently used. Over the past three decades, differential fear conditioning, a procedure that pairs one of two neutral stimuli with a mildly aversive unconditioned stimulus (UCS), has been widely used to study conditioned fear in humans. This conditioning procedure allows for the calculation of differential response scores, calculated as the average response to the conditioned stimulus (CS+) paired with the UCS minus the average response to the stimulus (CS-) not paired with the UCS. These scores reflect the amount of conditioned responding to the fear cue (CS+) during acquisition and extinction. Individuals with PTSD have been found to show larger differential SC 1114, corrugator EMG 11 and HR 1118 responses during acquisition and larger SC 111439 responses during extinction. Of particular relevance to the present work is the recent finding from a study of firefighter trainees that the magnitude of the differential corrugator EMG response during acquisition and extinction was significantly correlated with PTSS following exposure to a traumatic event 40.

To date, only a very few studies have addressed whether the aforementioned psychologic and psychophysiologic measures represent pre-trauma vulnerability markers or are consequences of developing PTSD/PTSS. From a practical standpoint, the identification of vulnerability markers for PTSD could have important practical implications related to screening and targeting early intervention for at-risk individuals working in high-risk occupations. From a theoretical standpoint, addressing whether or not psychologic and psychophysiologic features associated with PTSS represent pre-trauma or acquired traits has substantial implications for understanding the pathogenesis of PTSS and PTSD.

The work reported here assessed pre-trauma psychologic and psychophysiologic measures in order to evaluate their ability to predict the development of PTSD/PTSS severity following exposure to an intervening traumatic event. In order to increase the likelihood of finding participants who would be exposed to a traumatic event at some proximate point following pre-trauma assessment, we recruited firefighter and police trainees at the time of their initial training. The pre-trauma assessment included structured diagnostic interviews, standard psychometrics, and psychophysiologic reactivity to loud-tones and differential fear conditioning. Participants were closely followed for the occurrence of a traumatic event that met the DSM-IV PTSD Category A1 criterion during the course of their new professional duties. After such a traumatic event, the participant was invited to return for psychodiagnostic, psychometric and psychophysiologic assessments. As part of the post-trauma assessment, psychophysiologic reactivity to individually-tailored imagery scripts was measured as a non-subjective measure of emotional reactivity to the traumatic event.

The primary goal of the work reported here was to prospectively identify the best combination of psychologic and psychophysiologic variables that predicted PTSD/PTSS following exposure to a potentially traumatic event. Based upon the sparse but most relevant literature mentioned above, we hypothesized that individuals with higher PTSD/PTSS following exposure to a traumatic event would show the following pre-trauma traits: 1) greater conditionability (i.e., stronger acquisition and slower extinction of an aversively conditioned differential corrugator EMG response); 2) larger magnitude eyeblink EMG and SC responses to loud tones; 3) a higher mean SC level; 4) lower IQ; and 5) lower extraversion, higher neuroticism, and lower agreeableness scores on the Revised NEO Personality Inventory 41.

The current study, like those of Guthrie and Bryant 40 and Pole et al. 61, represents a truly prospective examination of potential pre-trauma physiologic, psychologic, and demographic predictors of PTSS in firefighters and police, as predictor variables were assessed prior to the occurrence of a traumatic event rather than retrospectively. The present work contributes to findings from previous studies, which have provided indirect evidence for the pre-existing versus acquired nature of psychologic and physiologic markers associated with PTSS 313638.

It is noteworthy that in the present study none of the 99 police and firefighters exposed to a traumatic event met full DSM-IV criteria for PTSD at the post-trauma assessment. The absence of clinical PTSD cases initially surprised us. When the study was originally designed we “conservatively” estimated that the PTSD rate would be about 17 % 46263. However, recent findings from prospective studies of police and firefighters suggest that a very low rate of PTSD following exposure to a traumatic event is characteristic of these populations. In their study of Australian firefighter recruits, Guthrie and Bryant 40 found that none of the 67 trauma-exposed firefighters met DSM-IV criteria for PTSD at the post-trauma assessment. Pole et al. 61 reported very low levels of PTSD symptom severities in police exposed to at least one traumatic event during the 1-year period following their initial training; only 1 of 138 police met criteria for full PTSD. Thus, the development of PTSD following exposure to a traumatic event appears to be very low in prospectively-studied samples of police and firefighters, likely lower than in the general population and perhaps reflecting a strong self-selection bias or protective benefits from professional training. One might assert that police and firefighters who are still actively engaged in their professions are simply reluctant to endorse symptoms of psychopathology. However, the fact that police and firefighters in the present study also showed normal levels of psychophysiological reactivity while recalling their traumatic events provides some objective support for the absence of significant PTSD symptoms.

The lack of police and firefighters who met full DSM-IV criteria for PTSD precluded our using the categorical diagnosis of PTSD as a post-trauma outcome and necessitated a shift in focus to pre-trauma predictors of PTSS. Despite the relatively low levels of PTSS, the combination of a higher pre-trauma depression score, lower estimated IQ and a larger differential corrugator EMG score during extinction trials of a fear-conditioning procedure significantly predicted higher PTSS. This finding accords well with accumulated evidence indicating that higher IQ appears to be a protective factor against developing PTSD 25 and the recent finding of Guthrie and Bryant 40 that a larger differential corrugator EMG response during extinction is predictive of a higher IES 64 score.

The process(es) underlying impaired extinction of corrugator EMG responses, lower IQ and higher pre-trauma depression symptoms appear to contribute to the pathogenesis of PTSS. Higher pre-trauma depression symptoms could reflect a propensity towards negative emotional affectivity 8. Corrugator EMG activity is typically interpreted as reflecting emotional valence 6566. The persistence of a larger differential corrugator EMG response following fear conditioning, could reflect a pre-trauma propensity towards the maintenance of negative emotional responses over time. Greater intellectual ability, as reflected in IQ, likely provides more extensive cognitive resources for managing negative emotional responses 24. Individuals with lower IQ would have more limited cognitive resources available for coping with negative emotions, which would contribute to their maintenance.

Stepwise logistic regression analysis identified the combination of lower estimated IQ and increased SC reactivity to loud tones as the best predictor of overall psychophysiologic reactivity during script-driven imagery. Guthrie and Bryant 37 and Pole et al. 61 previously reported that pre-trauma SC reactivity to loud tones (or noise) predicted subjective measures of PTSS. Their respective studies did not include a psychophysiologic outcome measure of PTSS. Taken together, findings from the studies of Guthrie and Bryant, Pole et al., and the present work suggest that pre-trauma individual differences in sympathetic nervous system reactivity, as manifest in larger SC responses to loud tones, can be used to predict the development of psychophysiologic and/or subjective PTSS. The inclusion of lower estimated IQ in the predictive model for post-trauma psychophysiologic reactivity suggests that more limited cognitive resources contribute to the maintenance of increased psychophysiological reactivity to trauma reminders, as well as contributing to the persistence of subjective PTSS.

Consideration of the SC and HR findings observed across studies using the loud-tone procedure, which provides an assessment of "unconditioned" responding, suggests that these measures may reflect different involvements of sympathetic and parasympathetic nervous system activity in the development and evolution of PTSS/PTSD. Skin conductance is primarily a measure of sympathetic nervous system activity 67, whereas HR is influenced by the parasympathetic, as well as sympathetic, nervous system. As noted above, there is accumulating evidence that increased SC reactivity to loud tones may be a pre-trauma predictor of PTSS, while there is very strong evidence that increased HR reactivity to loud tones is an acquired feature of PTSS/PTSD [e.g. 3638. Across studies of individuals with diagnosable PTSD, increased HR reactivity to loud tones has been reported in all but two studies, whereas increased SC reactivity has been less reliably obtained; interestingly, increased SC reactivity was observed in the two studies that failed to find increased HR reactivity (for review see 54). This mixed pattern of SC and HR findings in loud-tone studies suggests that, whereas increased sympathetic tone/activity underlies the increased SC reactivity, it is reduced parasympathetic, and not increased sympathetic, tone that is responsible for the heightened HR reactivity to loud tones. This leads to the interesting possibility that heightened sensitivity of the sympathetic nervous system represents a risk factor for, whereas reduced parasympathetic tone is a consequence of, developing PTSS/PTSD.

The present study also observed a significant relationship between the magnitude of the HR response to CS + stimuli during conditioning and the posterior probability score. This provides some limited support for a positive relationship between pre-trauma conditionability and psychophysiologic reactivity to trauma-related imagery. The fact that pre-trauma SC reactivity to loud tones and HR reactivity to CS + stimuli during conditioning predicted post-trauma psychophysiologic reactivity to script-driven imagery but not subjective PTSS (i.e., IES-R) scores raises the possibility that these positive relationships primarily reflect relationships between measures of psychophysiologic reactivity and not with PTSS per se. However, the likelihood that the observed positive relationships simply reflect concordance amongst measures of psychophysiologic reactivity is challenged by the fact that several important measures of psychophysiologic reactivity (e.g., SC orienting response, SC and corrugator EMG responses to CS + trials during conditioning, HR and eyeblink EMG responses to loud tones) did not show relationships with the posterior probability score.

Several limitations of the reported work should be kept in mind. The relatively low, i.e., normal, level of PTSS severity raises the possibility that the relationships observed in the present study between predictor and outcome variables may not extend to clinical levels of PTSD. However, it is also possible that the reported findings underestimate the actual presence and strengths of these relationships, given the restricted ranges of scores and impact this may have had on the correlation estimates. It is possible that the extended time periods between the traumatic event and follow-up assessment for some individuals may have adversely impacted the observed relationships. Although, when the logistic regression analysis was adjusted for the amount of time elapsed between the traumatic event and assessment, the results did not change. Our preference not to correct for multiple t-test comparisons is an additional limitation and caution should be exercised when interpreting the univariate results.

Finally, although the pre-trauma differential corrugator EMG score for extinction contributed to the prediction of IES-R scores in the expected direction, this was due in part to an unexpected pattern of responses in the Low IES-R group. More specifically, the Low IES-R group produced larger corrugator responses to the CS- than to the CS+, resulting in a negative differential score. We believe that the negative mean value for the Low IES-R group is reliable, given the substantial sample size for this subgroup (n = 67); however, the reason for larger corrugator responses to the CS-, compared to CS+, is not clear. As noted above, increased corrugator activity is generally thought to reflect increased negative emotional valence. It may be that when the associative link between CS + and UCS was broken during extinction, participants in the Low IES-R suspected that the experimenter might change the associative relationship such that the CS- would predict the UCS. Why this sort of cognitive processing would be more apparent in the Low IES-R, compared to High IES-R, group is not clear. Regardless of the underlying explanation for this curious pattern of responding to the CS + and CS- during extinction, it is important to acknowledge the potential predictive value of this measure.

The present study successfully identified several predictive relationships between pre-trauma psychological and psychophysiological measures and PTSS. Of particular interest, this study found that lower pre-trauma estimated IQ, a higher depression score and a larger differential corrugator EMG response following fear conditioning predicted higher IES-R scores, suggesting that the processes underlying these indices may contribute to the pathogenesis of PTSS. The combination of lower pre-trauma estimated IQ and increased SC reactivity to loud tones were found to predict greater psychophysiological reactivity (larger posterior probability score) to trauma-related cues, suggesting that processes related to these measures my contribute to the pathophysiology of PTSS. It is important to note that the levels of PTSS and general psychopathology were quite low in our sample of police and firefighters. Even so, the severity of post-trauma symptoms is dimensional and likely is normally distributed across the population. Therefore, it is possible that the predictors of PTSS identified in this study would generalize to a population with greater levels of symptomatology. It is plausible, if not likely, that the reported findings underestimate the actual presence and strengths of relationships between pre-trauma predictor variables and post-trauma outcome measures because of the limited variability in scores.

Prospective studies that identify predictors of emotional sequelae of traumatic events are particularly challenging because they necessitate extensive data collection from a sample wherein only a subset will be exposed to a traumatic event and a relatively small percentage of those exposed to trauma will ultimately develop PTSD. Despite the inherent challenges, these efforts are an important undertaking because of their potential to inform PTSD prevention and early treatment. A primary goal of prospective studies is to discover reliable markers that can aid in identifying individuals who are at highest risk for developing PTSS/PTSD and thereby allow more extensive resiliency training or early interventions to be targeted to those most likely to need them. Such interventions would be cost prohibitive if they were to be provided en masse and would be unnecessary for most individuals. Although significant challenges must be overcome when conducting prospective studies of PTSS/PTSD, it is our hope that future research will examine other high-risk populations, incorporate longer follow-up times and recruit larger samples. Such work would contribute substantively to the evolving literature aimed at identifying pre-trauma risk factors for PTSS/PTSD.

BDI, Beck Depression Inventory; CAPS, Clinician Administered PTSD Scale; CR, Conditioned Response; EMG, Electromyogram; HR, Heart Rate; IES-R, Impact of Event Scale-Revised; NEO-PI-R, NEO Personality Inventory-Revised; PTSD, Post-traumatic Stress Disorder; PTSS, Post-traumatic Stress Symptoms; SC, Skin Conductance; SCID, Structured Clinical Interview for DSM-IV; SCL-90-R GSI, Symptom Checklist 90-Revised; SSPQ, SCID Screen Patient Questionnaire; STAI, State-Trait Anxiety Inventory; UCR, Unconditioned Response; UCS, Unconditioned Stimulus; WAIS-R, Wechsler Adult Intelligence Scale-Revised.

The authors declare no competing interests.

SPO contributed to the conception and design of the study, oversaw the conduct of the study, oversaw management and scoring of the psychophysiological data, consulted on the data analyses and drafted the manuscript. NBL conducted all diagnostic interviews and assessments and helped edit the final manuscript. MLM recruited all police and firefighter participants, served as liaison with the training academies and followed participants for the occurrence of a traumatic event. SLP contributed to the drafting and editing of the manuscript. RKP contributed to the conception and design of the study and contributed to the drafting and editing of the manuscript. YC performed the statistical analyses and contributed to drafting and editing of the Results section of the manuscript. All authors read and approved the final manuscript.