Posttraumatic Stress Disorder Symptom Clusters as Predictors of Pain Interference in Burn Survivors: A Burn Model System National Database Study

Abstract

Individuals who experience burns are at higher risk of developing posttraumatic stress disorder and chronic pain. A synergistic relationship exists between posttraumatic stress disorder and chronic pain. We sought to evaluate the role of individual posttraumatic stress disorder symptom clusters as predictors of pain interference. We hypothesized that the hyperarousal and emotional numbing symptom clusters would be predictive of pain interference, even when accounting for the other two posttraumatic stress disorder symptom clusters, pain intensity, and other covariates. Multivariate linear regression analyses were completed using data from the Burn Model System National Database. A total of 439 adult participants had complete responses on self-report measures assessing posttraumatic stress disorder symptoms, pain intensity, and pain interference at 6-month after discharge and were included in analyses. Results indicate hyperarousal (B = .10, p = .03) and emotional numbing (B = .13, p = .01) posttraumatic stress disorder symptom clusters were each significantly associated with pain interference, even when accounting for pain intensity (B = .64, p < .001). Results highlight the importance of the emotional numbing and hyperarousal posttraumatic stress disorder symptom clusters in explaining pain interference. Findings suggest that when posttraumatic stress disorder symptoms or chronic pain are present, screening for and treating either condition may be warranted to reduce pain interference. Further, psychological interventions that target emotional numbing and hyperarousal posttraumatic stress disorder symptoms may be fruitful for promoting better coping with chronic pain and reducing pain interference.

INTRODUCTION

Individuals who experience burns are at higher risk of developing acute stress disorder (ASD) and posttraumatic stress disorder (PTSD).¹ The prevalence of PTSD in patients who have sustained a burn injury is thought to range from 2% to 40% between 3- and 6-month from the injury, and 9% to 45% in the year after injury¹ In addition to increased risk of psychological conditions such as ASD and PTSD, burn survivors are also at higher risk of acute and chronic pain.^2–4

High acute pain after sustaining burn injuries during the initial phase of treatment is common, and pain can last long into the rehabilitation phase and continue to be a problem for burn survivors years after their injury.² There is a broad range of prevalence of persistent pain following burn injury reported in the literature. Prevalence of pain postburn ranges from 30% to 52%.^3–11 Recently, the International Association for the Study of Pain (IASP) put forth criteria further distinguishing the development and classification of chronic pain in patients’ postsurgery and trauma, including burn injuries, and suggested that chronic pain is defined as pain that remained 3 months after healing.¹² Above pain intensity, pain interference is a central feature of the experience of chronic pain and resultant functional outcomes.¹³ Specifically, pain interference refers to the self-reported impact of pain on cognitive, social, emotional, physical, and recreational activities of patients, and can have a profound negative effect on burn survivors’ quality of life. In a survey study conducted by Dauber et al. (2002), 45% of burn survivors endorsed experiencing some persistent burn-related pain reported pain interference in their daily life.⁸ Pain postburn injury has been found to be a consistent predictor of unemployment.¹⁴ More recently, in a study performed with the larger study sample utilized in our study, pain interference was a significant predictor of later unemployment for those who endured moderate to severe burn injuries.¹⁵ Thus, pain and its interference on daily activities is an important outcome in relation to burn survivors’ longterm adjustment. Pain interference is also an important factor contributing to disability. The fear and avoidance model of pain suggests that fear of pain and avoidance of activities may provoke pain, followed by decreased activity, which fuels a cycle of disability.¹⁶

There is also reason to believe that other psychological symptoms and/or mechanisms may play a role in the development or maintenance of pain interference in burn survivors. For example, in both burn and nonburn populations, PTSD symptoms and chronic pain often co-occur.^6,17–21 High co-occurrence of PTSD and pain after multiple types of traumatic injuries suggests the context of the injury increases the probability of both PTSD symptoms and chronic pain. Theories exist about the shared vulnerability and mutual maintenance of PTSD and chronic pain that suggest there are both cumulative and independent contributions of PTSD symptoms and pain, as well as other underlying mechanisms perpetuating symptoms of both conditions.^22–24 Research has shown a synergistic relationship between PTSD and pain across injured patient populations^1,23 including individuals who have experienced burns.²⁰ In a sample of burn patients, Corry et al. found that PTSD and pain longitudinally were associated with psychosocial and physical disability.²⁰ Further, in a meta-analysis examining predictors of PTSD in patients with burns, severity of pain related to burn injury was found to be one of the significant predictors of development of PTSD symptoms.²⁵ Finally, research with active duty and returning veterans populations has shown both cross-sectional and longitudinal associations between PTSD symptoms and pain interference emphasizing the mutual maintenance which may occur between both PTSD and pain interference.^26–28

Given the complexity of the diagnostic criteria, individuals diagnosed with PTSD often present with significant symptom heterogeneity.²⁹ Past factor analyses favor a four-factor symptom cluster model of PTSD symptoms, splitting symptoms into the following clusters: reexperiencing, avoidance, hyperarousal, and emotional numbing (see Figure 1 for list of symptoms subsumed within each cluster).³⁰ It is thought that examining PTSD symptom clusters (as opposed to PTSD symptoms as one total factor) in relation to health outcomes can increase predictive utility^31,32 and therefore, may help to explain variance associated with outcomes, including pain interference.

Figure 1.

PTSD symptoms by cluster, in accordance with DSM-IV-TR criteria.

Previous research in nonburn patients suggests that PTSD symptom clusters have been shown to have differential associations with functional impairment related to pain. For example, one study that examined individual factors of PTSD in a sample of outpatient chronic pain patients found that hyperarousal symptoms were associated with pain intensity and pain-related disability.³³ Additionally, research in trauma exposed samples have demonstrated hyperarousal symptoms are associated with greater somatic and cognitive complaints in general.^34,35 Further, in studies that have evaluated a four-cluster PTSD model, emotional numbing has been shown to be associated with impaired role functioning a sample of patients experiencing pain and PTSD after motor vehicle accidents,³¹ and pain-related disability in a sample undergoing a lateral thoracotomy (Katz et al., 2009).³⁶ And in a large cross-sectional sample of patients with chronic musculoskeletal pain and PTSD, emotional numbing and hyperarousal symptoms were associated with both pain intensity and pain disability.³² Finally, in a sample of active duty service members were seeking care at a pain clinic, the emotional numbing domain accounted for most variance associated with pain interference; however, of note, PTSD symptom measurement was limited by the use of a brief screener and each cluster was measured by one to two items.²⁸

To better inform assessment, intervention, and promote positive adaptation, Corry, Klick, and Fauerbach²⁰ suggest that researchers evaluate the relationships between PTSD symptom clusters, rather than PTSD as one total diagnosis or score, with outcomes after burn injury. However, to our knowledge, no research has evaluated individual PTSD symptom clusters in relation to pain interference in burn injury populations. Thus, the aim of this study was to evaluate the relationship between the four PTSD symptom clusters (see Figure 1) and pain interference. We hypothesized that the hyperarousal symptoms and emotional numbing symptom clusters would be predictive of pain interference, even when accounting for pain intensity.

METHODS

Participants and Procedures

All data from the current study were collected as part of the Burn Model System Centers program (BMS) of research. The BMS, a federally funded multisite national research study evaluating outcomes of people with moderate-to-severe burn injuries, has kept a longitudinal database, since 1994. Eligibility requirements, data collected, and participating burn centers have changed over the years (https://burndata.washington.edu/about-bms). The measures used in the current study were added to data collection in 2015. Access to the database required authors to submit internal requests specifying the current study’s research aims and hypotheses.

As part of the larger BMS longitudinal database study participants complete surveys at discharge, 6- , 12-, 24-month and every 5 years postinjury. Participants are able to complete BMS surveys in-person during follow-up appointments, by phone, mailed paper-and-pencil surveys, or online. Further, some data are extracted from the medical record (eg, burn size, number of surgeries, and etiology of injury). Inclusion criteria for the larger BMS program include injuries requiring surgery for wound closure and at least one of the following: a burn to one or more critical areas (eg, face, hands, and feet); an electrical, high voltage, or lightning injury; or burn sizes with a total body surface area (TBSA) ≥10% for older adults 65 years and ≥20% for people between the ages of 0 and 64 at the time of injury. For inclusion in the current study’s analyses, participants were aged 18 years or older at the time of follow-up data collection and needed complete responses on the current study’s measures of interest (eg, PTSD symptoms, pain interference, and pain intensity). Thus, individuals enrolled prior to 2015 were also excluded due to the addition of the measures of interest in data collection in 2015. Data from the current study utilized data from the second timepoint (6-month after discharge). Study procedures were approved by institutional review boards at all participating BMS institutions. Database structure, enrollment, follow-up strategies, and data verification processes have been previously described in detail.^37,38

MEASURES

Demographic information was collected at initial enrollment and included sex, age, race and ethnicity. Injury characteristics including size of burn (percentage of total body surface area burned: % TBSA) and length of hospital admission were assessed at the time of discharge using medical record abstraction.

PostTraumatic Stress Disorder Checklist-Civilian (PCL-C)

The PostTraumatic Stress Disorder Checklist-Civilian (PCL-C)³⁹ was developed in accordance with DSM-IV criteria to screen individuals for PTSD, monitor symptom change due to interventions, and to make a provisional PTSD diagnosis. The measure is a reliable and valid measure of PTSD symptoms and it is widely used. For the current study, items were summed in accordance with the four-cluster model of PTSD symptoms.⁴⁰ Resultant scores were for the following PTSD symptom clusters: Avoidance, Reexperiencing, Emotional numbing, and Hyperarousal. Total scores for the full measure can range between 17 and 85, with higher scores reflecting greater symptom severity. However, scores for each cluster vary, and can range between 5 and 25 for the ReExperiencing, Emotional Numbing, and Hyperarousal clusters, and between 2 and 10 for the Avoidance cluster. The threshold for a clinically elevated PTSD symptoms is considered to be 44; however, it is also possible to receive a diagnosis with a total score of 29 given the requirements for each cluster.³⁹

Pain Intensity

Participants’ current pain intensity was measured using a single item from the Patient Reported Outcomes Measurement Information System (PROMIS-29^®).⁴¹ The item is a legacy pain intensity item commonly used in the literature which asks patients to rate their pain on a 0 to 10 point scale. The item specifically asks participants to rate their “pain on average in the last 7 days” on an 11-point scale from 0 (no pain) to 10 (worst imaginable pain). Pain scores were not categorized but were included in the regression model on the 0 to 10 point scale.

Pain Interference

Pain interference was measured using the four-item pain interference subscale embedded as part of the PROMIS-29 profile v2.0 measure.⁴¹ Participants reported the level of interference pain has had on functional activity, such as overall day-to-day function, as well as in specific domains such as work around the home, social activities, and household chores. Consistent with standardized PROMIS short form scoring, summary scores for the pain-interference items were converted to T-scores (see healthmeasures.net for user guides and scoring information). Scores on the PROMIS pain interference domain are centered on the general U.S. population (M = 50, SD = 10), and higher scores reflect worse pain interference.⁴²

Analyses

Study analyses were conducted using Stata.⁴³ We compared study participants to participants excluded for incomplete data on demographic variables to evaluate the representativeness of our sample. We used t-tests for continuous data (age, burn size, and days until discharge) and chi-square tests for categorical variables (sex, race, and ethnicity). Consistent with past literature we included age, sex, days in the hospital, and percentage of total burn surface area (TBSA) burn size as covariates in our current study.²⁰ Linear regression models were fit, using robust variance estimators to account for heteroskedasticity. The linear regression model examining predictors of pain interference included the four PTSD symptom clusters (Avoidance, Reexperiencing, Emotional Numbing, and Hyperarousal), pain intensity, and our study covariates (age, sex, days in the hospital, and %TBSA burn size). All variables were included in one a priori defined model regardless of statistical significance, and standardized betas were calculated to evaluate the relative strength of the associations of the symptom clusters and other variables with pain interference.

RESULTS

Sample Characteristics

As shown in the study sample flowchart (Figure 2), of the 740 individuals eligible for this study, 439 respondents had complete data on variables of interest and were included in analyses. The large majority of those excluded (n = 301) were due to missing data at the 6-month follow-up timepoint (n = 187), where only medical record data were collected. The remainder, either only partially completed survey assessments (n = 52), were lost to follow-up (n = 50), refused assessment (n = 3), or were otherwise unable to respond due to health issues (n = 8). Overall, our sample was mostly white (82%), non-Hispanic (18%), and male (69%). The average age of participants was 47.2 (SD = 16.1), with the youngest being 18-year-old and the oldest 88-year-old. Participants in the study sample had a mean TBSA burn size of 18% (SD = 18.6), with the lowest TBSA in our sample being 0.1% and the largest burn size being 85%. Participants had a mean hospital stay of 27.2 days (SD = 26.2), with the shortest hospital stay of 1 day and the longest being 165 days. See Table 1 for demographics. Participants with complete data who were included in the analyses differed from those excluded from analyses (n = 301) for hospital length of stay and ethnicity. Specifically, respondents excluded from analyses had a higher average length of stay and a greater percentage identified as Non-Hispanic/Latino. Overall, the study sample reported slightly higher (M = 52.56) average levels of pain interference than the U.S. general population.) Respondents endorsed low to subthreshold levels of PTSD symptoms on average (M = 28.80, SD = 4.38). Refer to Table 2 for means, standard deviations, and ranges of our other study variables. Almost all main study variables of interest were intercorrelated. All nonsignificant correlations included either age or sex. For significant correlations, r’s ranged from 0.10 to 0.78.

Figure 2.

Recruitment and participation flowchart for Burn Model Systems study and current study.

Table 1.

Demographic descriptive statistics and comparisons between responders and nonresponders

	Responders	Nonresponders
	n = 439	N = 301
Variable	Mean ± SD	Mean ± SD
	or	or
	n (%)	n (%)	t-test or chi2
Age (years)	47.2 ± 16.1	46.5 ± 15.7	t = −0.56
TBSA Burned (%)	17.9 ± 18.6	18.8 ± 20.4	t = 0.67
Days Until Discharge	27.2 ± 26.2	34.8 ± 46.5	t = 2.82**
Sex
Female	134 (31%)	104 (35%)	chi2 = 1.32
Male	305 (69%)	197 (65%)
Race (n = 426)
Black	44 (10%)	36 (13%)	chi2 = 2.59
Caucasian	355 (83%)	225 (79%)
Other	27 (6%)	25 (9%)
Ethnicity (n = 423)
Non-Hispanic/Latino	345 (82%)	254 (87%)	chi2 = 4.18*
Hispanic/Latino	78 (18%)	37 (13%)

Note: TBSA, total burn surface area.

^*p < .05,

^**p < .01.

Table 2.

Descriptive statistics and ranges for main study variables

Variable	M	SD	Minimum	Maximum
PTSD Symptom Total	28.80	4.38	17	85
PTSD Reexperiencing	8.35	4.82	5	25
PTSD Hyperarousal	9.19	4.64	5	25
PTSD Avoidance	3.30	2.07	2	10
PTSD Emotional Numbing	7.96	4.38	5	25
Pain Intensity	3.19	2.82	0	10
Pain Interference	52.56	10.59	41.60	75.60

Note: PTSD, posttraumatic stress disorder.

Linear Regression Model

See Table 3 for results. Participants’ hyperarousal (B = .10, p = .03) and emotional numbing (B = .13, p = .01) symptom PTSD clusters were both significantly associated with greater pain interference. With higher reports of symptoms from these clusters being associated with higher pain interference. The other two symptom clusters were not significantly associated with pain interference. Pain intensity was also a significant predictor of pain interference (B = .64, p < .001), with greater pain intensity being associated with higher pain interference. None of the other covariates included in the model were significant predictors of pain interference. The final model accounted for 61% of the variance in pain interference (F = 85.52, p < .001, R² = .61).

Table 3.

Linear regression model examining associations with pain interference (n = 439)

Predictor Variable	B	SE	t	F	R2
Age	.04	.02	1.15	85.52***	.61
Sex	−.02	.75	−.67
Days Until Discharge	.04	.02	.99
Total Burn Surface Area	.01	.02	.33
Pain Intensity	.64***	.16	15.14
PTSD Reexperiencing	.07	.14	1.07
PTSD Hyperarousal	.10*	.11	2.20
PTSD Avoidance	−.05	.28	−.94
PTSD Emotional Numbing	.13**	.11	2.70

Note: B, standardized beta coefficient; SE, standard error; PTSD = posttraumatic stress disorder.

^*p < .05;

^**p < .01;

^***p < .001.

DISCUSSION

Our study evaluated PTSD symptom clusters as predictors of pain interference that is chronic in nature. Approximately 6 months after discharge burn survivors’ reports of hyperarousal and emotional numbing PTSD symptom clusters were significantly associated with greater pain interference. This was the case even after adjusting for reexperiencing and avoidance PTSD clusters and pain intensity. Given this timeline, pain reported by survivors can be classified as chronic pain. These findings support that specific PTSD symptom clusters account for unique variance associated with pain interference.

Out of the PTSD symptom clusters, the emotional numbing cluster, which includes anhedonia, feelings of detachment, restricted emotional expression, and low positive emotion, was the cluster most highly associated with pain interference. This finding is consistent with prior findings conducted with nonburn patients demonstrating the influence of emotional numbing on impaired role function and disability.^31,36 One explanation is that the anhedonia¹ associated with emotional numbing symptoms may be related to greater pain interference as the result of lower activity levels, which play a role in the pain-disability cycle. Past research with burn survivors has demonstrated the interaction between pain and depression in predicting physical functioning.⁴⁴

The emotional numbing cluster also includes symptoms that are thought to be reflective of dissociative symptoms (eg, restricted emotional expression, amnesia, and detachment). Also referred to as “emotional anesthesia,” this symptom cluster is implicated in the release of endogenous opioids during the fear responses.^45,46 Emotional numbing symptoms can be thought of as a deactivating nervous system response to high stress and the inability to regulate intense emotion and cope with demands. This cluster of symptoms can also be considered a form of avoidance or escape in response to distress, which may result in decreased engagement in activities.³⁷ Avoidance is thought to be a shared mechanism perpetuating both PTSD and pain.³⁸ In general, the emotional numbing symptom cluster is conceptualized as being reflective of PTSD severity, with greater occupational and social impairment.^47–51

Study results also revealed that the hyperarousal symptom cluster was a significant predictor of pain interference. One explanation for this is the association between hyperarousal symptoms, sleep problems, and somatic complaints, which are in turn thought to be more predictive of greater pain problems and increased physical arousal. Hyperarousal symptoms are related to higher sympathetic activity (eg, increased heart rate) and less parasympathetic control. Patients with PTSD symptoms may experience increased muscle tension as well as increased heart rate, skin conductance, and blood pressure (see review by Pole).⁵² Wall and Melzack⁵³ posit that the link between emotional arousal and pain can be explained by increased activity in the central nervous system, autonomic nervous system, and musculoskeletal system. Another component of hyperarousal worth consideration includes the attentional component in which patients may become hyperaware and hypervigilant in response to their physiological arousal, perceiving the arousal as threat, which can lead to overmonitoring symptoms of pain, or subsequent restriction of activity out of fear of exacerbating pain. In fact, in a study of burn patients, Sgroi et al.⁵⁴ found that hyperarousal symptoms were strongly associated with fear-avoidance beliefs underlying pain and occupational impairment, suggesting cognitive behavioral mechanisms that may underly the associations found in the current study.

Taken together, there is also reason to believe that both hyperarousal and emotional numbing may synergistically work together or have other underlying cognitive-behavioral mechanisms that relate to pain outcomes. For instance, in a large cross-sectional sample of patients with musculoskeletal pain and PTSD, Lopez-Martinez et al. found that of the four PTSD symptom clusters, hyperarousal and emotional numbing were predictive of both pain intensity and pain disability via participants’ level of anxiety sensitivity, fear of pain, and catastrophizing.⁵⁵ Further, theories on posttraumatic numbing also postulate that chronic hyperarousal may lead to emotional numbing⁴¹ and correlational findings have supported this link.^31,56,57 Although it is believed emotional numbing symptoms may be protective against harm and acute pain in the short-term, it is probable that over time it becomes maladaptive.

Pertinent to treating burn survivors, these results highlight the relevance of assessing psychological symptoms in evaluating pain interference. These findings are relevant for clinicians. Specifically, when treating patients with chronic pain reporting high functional interference, clinicians should also consider screening for PTSD symptoms. Similarly, for patients presenting with PTSD symptoms, it may be helpful to assess for health outcomes such as chronic pain and pain interference with daily activities. Recent work by Stockly et al. has helped to institute a risk scoring system to stratify those at risk of PTSD following burn injury.⁵⁸

Although more research is needed on the role of PTSD symptoms and chronic pain in burn survivors longitudinally, early, and effective treatment of PTSD symptoms is an important aspect of pain management for burn survivors. There are several evidence-based individual psychotherapies recommended as first line treatments by different clinical practice guidelines, such as Cognitive Processing Therapy, Prolonged Exposure, and Eye Movement Desensitization Reprocessing Therapy.^2,59

Given our results, modular treatment of hyperarousal and managing related cognitive and somatic complaints may be useful for individuals experiencing increased physiological arousal as a means of managing their pain interference. Such interventions to increase parasympathetic arousal might include relaxation training, biofeedback, diaphragmatic breathing, and progressive muscle relaxation.

Although our study hypotheses were supported by our data, our study has potential limitations. First, our current study findings are cross-sectional; therefore, limiting our ability to draw any inferences about causality between our variables of interest. Further, our methods of assessment of study variables were limited to self-report measures, which may affect accuracy of reporting, and may be affected by other reporting biases. Additionally, the PTSD measure utilized in the current study identified symptoms based on DSM-IV-TR criteria, rather than the current DSM-5 criteria.⁶⁰ Further, we do not have measures of ASD or PTSD symptoms at baseline or closer to injury and initial hospitalization. History of prior trauma exposure, depression, and other factors would be helpful to consider in future analyses as well.

Future research should evaluate longitudinally whether symptom clusters are predictive of pain interference. Future research with burn survivors should evaluate other broad or overlapping theoretical constructs that may also account for the link between PTSD symptoms and chronic pain, such as fear and avoidance beliefs, anxiety sensitivity, catastrophizing, coping strategies, and other mental health and somatic symptoms. Better understanding of mutual mechanisms underlying both pain and PTSD symptoms in burn survivors is needed to inform treatment aimed at alleviating and to help bolster patients’ function and quality of life postinjury.