Relationships among childhood trauma, central sensitization, and opioid-related risks in a community sample of adults with chronic pain on opioids

Abstract

Objectives

This study examined the relationships among childhood trauma, central sensitization, and pain-, mood-, and opioid risk-related outcomes among adults with chronic pain (CP) on long-term opioid therapy (LTOT). We hypothesized that Adverse Childhood Experiences (ACEs) would relate to poorer outcomes, and that central sensitization would mediate the relationships between ACEs and outcomes.

Methods

Adults with CP, actively prescribed opioids, were recruited via ResearchMatch.org and completed a cross-sectional electronic survey. Participants completed the Adverse Childhood Experiences (ACEs) questionnaire, the Central Sensitization Inventory (CSI), the Brief Pain Inventory (BPI), the Center for Epidemiological Studies Depression scale (CES-D-10), the Generalized Anxiety Disorders 7-item questionnaire (GAD-7), the Rapid Opioid Dependence Screen (RODS), and other substance use items. Models were estimated examining main effects of ACEs and potential mediating role of CSI between ACEs and outcome variables.

Results

ACEs positively predicted higher CSI scores (β = 0.434, P<0.001), opioid dependence symptoms (β = 0.199, P<0.05), lifetime polysubstance use (β = 0.367, P<0.001), depressive symptoms (β = 0.262, P<0.001), and anxiety symptoms (β = 0.251, P = 001). As a mediator, CSI symptoms significantly mediated the association between ACEs and pain interference (LCL = 0.124, UCL = 0.303), ACEs and depressive symptoms (LCL = 0.227, UCL = 0.438), and ACEs and anxiety symptoms (LCL = 0.212, UCL = 0.387).

Conclusions

Among a sample of adults with a variety of CP conditions on LTOT, ACEs emerged as a risk factor for some substance use and mood outcomes, though inconsistent effects suggest other influential factors. CSI mediated the relationship between ACEs and pain interference, depression, and anxiety, indicating childhood trauma may be especially impactful when central sensitization is higher.

This cross-sectional study examined the relationships among childhood trauma (Adverse Childhood Experiences; ACEs), central sensitization, and pain-, mood-, and opioid risk-related outcomes among adults with chronic pain (CP) using opioids. Findings demonstrated that ACEs were a risk factor for higher central sensitization and some substance use and mood-related outcomes. Further, central sensitization mediated the association between ACEs and depressive symptoms, between ACEs and anxiety symptoms, and between ACEs and pain-related activity interference, highlighting the complex nature of the relationship between ACEs and later outcomes and the potential role of stress sensitization in CP adaptation.

Introduction

The dual crises of chronic pain (CP) and opioid overdoses are widespread, interconnected public health challenges. CP affects over 50 million American adults¹ and results in significant costs at both the individual and societal levels. Meanwhile, the opioid epidemic has resulted in devasting rates of fatal overdoses² and has disproportionately affected individuals with CP.³ Overprescribing of opioid pain medications fueled the first wave of the opioid crisis,⁴ and although synthetic opioids like fentanyl now drive most overdoses,⁵ 13 million Americans remain on long-term opioid therapy (LTOT),⁶ which is associated with increased risk of overdose and drug-related deaths.⁷ In particular, comorbid substance use among individuals on opioids for pain further heightens the risk of adverse outcomes, including overdose.⁸ To optimize prevention and treatment approaches, it is vital to understand factors that impact both pain outcomes and risk behaviors—including substance use—among individuals with CP on LTOT.

Childhood trauma has been identified as a key risk factor for development of both pain and substance use disorders.⁹ A recent meta-analysis found that individuals with at least one adverse childhood experience (ACE), such as abuse, neglect, or household dysfunction, were 45% more likely to have CP in adulthood, and an increasing number of ACEs correlated with higher CP prevalence and pain-related disability.¹⁰ Among individuals with CP on LTOT, childhood trauma is associated with higher use of prescription opioids,^11–14 opioid misuse,¹⁵ and opioid use disorder (OUD).¹⁶ While mood disorders like depression and anxiety contribute, they do not fully explain these associations.^17,18 Emerging evidence points to central sensitization—a pathophysiological process involving heightened nervous system reactivity¹⁹—as another key mechanism. Central sensitization, which is characterized by increased pain sensitivity, including dynamic tactile allodynia, heightened response to pressure or punctate stimuli, prolonged sensations after stimulation, and intensified pain with repeated stimuli,²⁰ may bridge early life trauma and maladaptive pain processing. Preclinical and clinical studies show that early life stress (childhood adversity) contributes to neurobiological changes and heightened sensitivity of the stress response,²¹ which may increase susceptibility both to CP disorders, especially those characterized by central sensitization, as well as substance use. For example, conditions such as fibromyalgia, where central sensitization is the most prominent feature, are more common in individuals with childhood adversity.²² In adolescent chronic pain samples, higher levels of central sensitization have been correlated with childhood abuse.²³ Outside of chronic pain samples, central sensitization has been associated with elevated risk of opioid use disorder onset and relapse²⁴ as well as poorer outcomes among individuals with alcohol use disorder (AUD), including elevated alcohol use to cope with pain.²⁵ Another study found that in a sample with both pain and OUD, fibromyalgia symptomatology predicted greater risk for pain-related OUD relapse,²⁵ which, to our knowledge, was the only study to concurrently evaluate childhood trauma and central sensitization as potential related risk factors for substance use outcomes in CP.

While prior work has explored links between childhood trauma, pain, and substance use, few studies have jointly examined ACEs, central sensitization, and opioid-related risks in LTOT patients who do not carry an OUD diagnosis. This distinction is clinically important, as LTOT patients can be challenging to treat given fewer treatment guidelines for complex persistent dependence²⁶ versus OUD, lower levels of motivation for nonpharmacological pain treatments once on LTOT^27,28 and comorbid mental health and addiction risk concerns.^11,29–31 Further, if these individuals co-use other substances in addition to prescribed opioids, risk for anxiety, depression, and overdose are elevated.^32,33 Clarifying mechanisms of risk could inform tailored prevention and treatment approaches to improve safety and overall outcomes. We hypothesized that among a community sample of patients with CP on LTOT, ACEs would be related to increased opioid-related risks and co-use of other substances to manage pain, as well as poorer pain and mood-related outcomes, including higher depression, anxiety, pain severity, and pain interference. Further, we explored central sensitization as a more proximal indicator that may help explaining these relationships and hypothesized that central sensitization would mediate the relationships between ACEs and pain, mood, and substance use indicators.

Methods

Participants and procedures

Participants were recruited nationally using ResearchMatch.org, a nonprofit research platform sponsored by the National Institutes of Health (NIH) that connects researchers with volunteers interested in participating in research studies. ResearchMatch has over 120,000 volunteers from all 50 states in the United States. Volunteers were recruited who met the following criteria: 1) having a CP condition and 2) taking an opioid medication. Volunteers who met these initial criteria and indicated interest then completed a screening survey on REDCap to determine whether they met all eligibility criteria including: 1) being 18 or older; 2) living in the United States and being English-speaking; 3) experiencing CP defined as pain on at least half the days lasting for 3 months or longer; and 4) currently taking a prescription opioid medication (a list of opioid medications was provided).

Eligible participants completed a battery of electronic, self-report questionnaires in REDCap. They responded to a series of items assessing demographic and medical characteristics including gender, age, race, education, employment status, household income, marital status, medical care, and mental health diagnoses. Pain-specific variables were assessed including pain diagnoses, number of pain areas in the body (selected from a list of body areas), pain duration, and use of pain medications and other medical treatments. Several attention items were included, e.g., “answer 3 for this question,” to verify valid and attentive responding. Upon completing the survey, participants received a $15 electronic gift card. The study was approved by the Virginia Commonwealth University Institutional Review Board.

Measures

Adverse childhood experiences (ACEs)

History of experiencing adverse events in childhood was measured using the Adverse Childhood Experiences questionnaire.^34–36 This 10-item measure asks respondents to select a response of “yes” or “no” regarding experiencing a number of events including physical abuse, sexual abuse, neglect, or household dysfunction prior to their 18^th birthday. Total scores are calculated as a count (sum) of endorsed events and range from 0 to 10. The ACEs questionnaire has been used with a variety of populations, and a score of 4 or above has been linked to significantly poorer long-term health outcomes.³⁷

Central sensitization

Central sensitization was assessed using the Central Sensitization Inventory (CSI).³⁸ For the purposes of the present study, Part A of the CSI was used to determine the frequency of symptoms commonly associated with syndromes of central sensitivity. Respondents are asked to rate each of the 25 items on a 5-point scale ranging from 0 (“Never”) to 4 (“Always”). Total scores range from 0 to 100, with higher scores indicating greater symptom burden. A systematic review of psychometric properties of the CSI found that the measure has demonstrated good-to-excellent internal consistency, reliability, structural validity, cross-cultural validity, responsiveness, content validity, and criterion validity.³⁹

Opioid and substance use

Recent opioid use was determined by participants’ responses to the question, “Have you taken a prescription opioid medication in the past 2 years for at least 7 days? Prescription opioid medications are pain medication such as [list provided].” Current opioid use was determined by participants’ responses to the item, “Do you currently take a prescription opioid medication for your pain? Prescription opioid medications are pain medications such as [list provided].” Morphine equivalent daily dose (MEDD) was determined based on participants’ self-reported medication type(s) and dose(s), if sufficient information was provided to do so, using a publicly available conversion calculator that applies standard conversion factors aligned with CDC guidelines.⁴⁰

Likelihood of opioid dependence was measured using the Rapid Opioid Dependence Screen (RODS).⁴¹ This 8-item measure was designed to aid in quick screening for opioid dependence in both clinical and research settings. The first item is used to determine respondents’ history of opioid use in any context and is excluded from scoring. The remaining seven items ask respondents to select a response of “yes” or “no” regarding the presence of symptoms of opioid dependence (e.g., “Did the idea of missing a dose ever make you anxious or worried?”) and are summed to achieve a total score. Scores of 3 or greater have been associated with excellent sensitivity, specificity, positive predictive value, and negative predictive value for opioid dependence according to DSM-IV criteria.⁴¹

Indicators of polysubstance use were assessed using binary (“yes/no”) questions endorsing use of various substances. Lifetime use (“In your entire life, have you EVER used…”) and current/recent use (“In the past 3 months, have you used…”) of substances were assessed across 9 substance classes: Tobacco, alcohol, cannabis, cocaine, amphetamine, inhalants, sedatives or sleeping pills, hallucinogens, or non-prescribed opioids. Respondents were instructed to only consider non-medical, non-prescribed use for all substance categories. Two polysubstance use variables (lifetime polysubstance use and current polysubstance use) were computed as count variables reflecting the number of affirmative responses to these yes/no questions (maximum total score = 9). For any current/recent substance use endorsed, participants also indicated whether they use the substance for pain, with the item “To what degree do you use [substance endorsed] for pain” with three response options indicating whether they use the substance for pain relief: Values for this variable were coded as 0 (“not at all”), 1(“occasionally”), and 2 (“usually or always”).

Depression and anxiety

Depressive symptoms were measured using the 10-item version of the Center for Epidemiological Studies Depression scale (CES-D-10).⁴² The CES-D-10 asks respondents to rate how often they have experienced a list of symptoms over the past week. Response options range from 0 “rarely or none of the time (less than one day)” to 4 “all of the time (5–7 days),” with total scores ranging from 0 to 40. The CES-D-10 has been widely used with a variety of populations and has demonstrated strong reliability and validity in general samples.⁴³

The Generalized Anxiety Disorders 7-item questionnaire (GAD-7)⁴⁴ was used to measure anxiety symptoms. Respondents select how often they have been bothered by a given symptom over the past two weeks, with response options ranging from 0 “not at all” to 3 “nearly every day.” Total scores range from 0 to 21, with higher scores indicating greater anxiety symptoms. A cutoff score of 10 or above is frequently used as an indicator of clinically significant anxiety symptoms. The GAD-7 has well-established reliability and validity for use in the general population.⁴⁵

Pain severity and interference

Pain severity and interference were measured using the Brief Pain Inventory (BPI),⁴⁵ a widely-used, well-validated and reliable measure of pain and pain interference in many chronic pain populations (α = 0.80–0.92).^46,47 Respondents rated their worst (in past 24 hours), least (in past 24 hours), average, and current pain intensity on a scale of 0 (no pain) to 10 (pain as bad as you can imagine). Respondents then rated the degree to which pain had interfered in seven life areas (general activity, mood, walking ability, normal work, relations with other people, sleep, and enjoyment of life) in the past 24 hours. Pain interference was rated on a scale of 0 (does not interfere) to 10 (completely interferes).

Analytic plan

All statistical models were estimated using Mplus, version 6.12. Data were examined for statistical assumptions prior to all analyses, and all assumptions, including normal distribution of data, were met for all continuous variables used in ordinary least squares regression in the current analysis. One exception to this analytic approach was our polysubstance use variables, which were calculated as count variables and specified as categorical variables in the analysis; models including these polysubstance use variables were instead estimated using the weighted least squares mean and variance-adjusted (WLSMV) estimator, which is preferred for categorical or count variables due to its strong ability in handling non-normal distributions and ordinal outcomes, especially when sample sizes are moderate. Missing data patterns were assessed via Little’s MCAR test, and estimation maximization was used to impute data missing completely at random (<5% of all data). Main effects of adverse childhood experiences were estimated on several endogenous variables: MEDD, opioid dependence symptoms, history of polysubstance use, current polysubstance use, polysubstance use to manage pain, pain severity, pain-related interference, depressive symptoms, and anxiety symptoms. Cross-sectional models were estimated examining central sensitization symptoms as potential mediating factors between ACEs and each outcome variable. Mediation analysis was conducted only in cases where the exogenous variable (ACEs) or the mediator (CSI scores) showed a significant association with the outcome. This decision was made to preserve statistical power and focus interpretation on more robust pathways. All models were estimated with 1000-draw bootstrapped standard errors and mediated effects were estimated with 95% bootstrapped confidence intervals (LCL = lower confidence limit; UCL = upper confidence limit); these confidence intervals reflect the range of values within which the true parameter for an estimated mediated effect is likely to fall with 95% certainty. Age and sex were included as covariates in all models.

In order to adjust for multiple comparisons, a Benjamini-Hochberg False Discovery Rate (FDR) correction was applied, which establishes statistical significance by rank-ordering the statistical significance values of all estimated models and comparing these rank-ordered P values to adjusted P value cutoffs based on the total number of estimated models.⁴⁷ Proportion of variance accounted for in each outcome variable (r²) reflect only the main effect of a single exogenous variable (ACE or CSI scores) excluding covariates or other predictors to aid in interpretation of the relative size of each hypothesized relationship. This step was taken to allow the relative contributions of ACE and CSI scores to be compared directly and to avoid potential overestimation of the size of the hypothesized relationships⁴⁸ (as would have been possible if the models used to calculate r² values also included covariates).

Results

The current sample was comprised of 157 respondents. All demographic and clinical characteristics of the study sample can be found in Table 1. Regarding gender, the majority of the sample (n = 124, 79.0%) self-identified as female, 28 respondents (17.8%) self-identified as male and 4 participants (2.5%) self-identified as a gender other than male or female. Regarding self-reported racial identity, the most commonly endorsed racial categories were Caucasian (n = 130, 82.8%), African American (n = 22, 14.0%), “other” racial identity (n = 8, 5.0%), and multiracial (n = 7, 5.0%). Participants were allowed to endorse greater than one racial category. Mean age was 51.6 years (SD = 13.0). On the medical/pain history items, participants endorsed having received an average of 6.3 different pain diagnoses, endorsed 6.3 areas of bodily pain, and the majority of respondents (n = 104, 66.2%) reported having pain for more than 10 years. More than half of the sample endorsed having a diagnosis of depression (n = 91, 58.0%) or anxiety (n = 83, 52.9%). Regarding ACEs, 65 participants (41.4% of the sample) endorsed 4 or more ACEs. Tables 2 and 3 provide descriptive data on the variables of interest.

Table 1.

Sample characteristics (N = 157).

	n (%)	M (SD)
Gender
Woman	124 (79.0)
Man	28 (17.8)
Other	4 (2.5)
Age		51.6 (13.0)
Race
Caucasian	130 (82.8)
African American	22 (14.0)
Multiracial	7 (4.5)
Other	8 (5.0)
Number of pain diagnoses		6.3 (3.0)
Number of pain areas		6.3 (3.3)
Pain duration
<1 year	4 (2.5)
1–5 years	18 (11.5)
5–10 years	31 (19.7)
10 years +	104 (66.2)
Primary pain diagnosis type
Back pain	46 (29.2)
Arthritis	26 (16.6)
Fibromyalgia	16 (10.2)
Neuropathic	11 (7.0)
Musculoskeletal	9 (5.7)
Post-injury	9 (5.7)
CRPS	8 (5.1)
Headache	5 (3.2)
Other	27 (17.2)
Pain severity average (0-10)		5.8 (1.6)
Pain interference average (0-10)		6.2 (2.0)
Mental health diagnosis
None	52 (33.1)
Depression	91 (58.0)
Anxiety	83 (52.9)
Other	36 (22.9)

Table 2.

Descriptive statistics of key study variables.

Variable	Mean (SD, Range) or Frequency (%)
ACEs	3.4 (2.7, 0-9)
ACEs: 0-1	55 (35.0%)
ACEs: 2-3	37 (23.6%)
ACEs: 4 or more	65 (41.4%)
CSI	47.6 (18.8, 1–92)
RODS	1.4 (1.7, 0-7)
RODS: 3 or greater (above cutoff)	33 (21.0%)
Current polysubstance use (sum)	1.0 (1.2, 0-6)
Lifetime polysubstance use (sum)	3.3 (2.4, 0–9)
CES-D	13.9 (6.6, 1–30)
GAD-7	7.2 (5.6, 0–21)
BPI Pain Severity	5.8 (1.6, 2.5–10)
BPI Pain Interference	6.2 (2.0, 1–10)

Table 3.

Rates of endorsed substance use by substance type.

	Lifetime use (n, %)	Current use (n, %)	Use for pain (n, %)
Tobacco	No: 70 (44.6%) Yes: 85 (54.1%) Missing: 2 (1.3%)	No: 120 (76.4%) Yes: 36 (22.9%) Missing: 1 (0.6%)	Never: 149 (94.9%) Occasionally: 7 (4.5%) Usually/Always: 1 (0.6%)
Alcohol	No: 32 (20.4%) Yes: 125 (79.6%) Missing: 0 (0%)	No: 100 (63.7%) Yes: 56 (35.7%) Missing: 1 (0.6%)	Never: 143 (91.1%) Occasionally: 13 (8.3%) Usually/Always: 1 (0.6%)
Cannabis	No: 64 (40.8%) Yes: 93 (59.2%) Missing: 0 (0%)	No: 117 (74.5%) Yes: 37 (23.6%) Missing: 3 (1.9%)	Never: 126 (80.3%) Occasionally: 16 (10.2%) Usually/Always: 15 (9.6%)
Cocaine	No: 120 (76.4%) Yes: 33 (21.0%) Missing: 4 (2.5%)	No: 154 (98.1%) Yes: 1 (0.6%) Missing: 2 (1.3%)	Never: 157 (100%) Occasionally: 0 (0%) Usually/Always: 0 (0%)
Amphetamine	No: 112 (71.3%) Yes: 43 (27.4%) Missing: 2 (1.3%)	No: 148 (94.3%) Yes: 8 (5.1%) Missing: 1 (0.6%)	Never: 152 (96.8%) Occasionally: 3 (1.9%) Usually/Always: 2 (1.3%)
Inhalants	No: 142 (90.4%) Yes: 10 (6.4%) Missing: 5 (3.2%)	No: 153 (97.5%) Yes: 2 (1.3%) Missing: 2 (1.3%)	Never: 155 (98.7%) Occasionally: 1 (0.6%) Usually/Always: 1 (0.6%)
Sedatives	No: 83 (52.9%) Yes: 70 (44.6%) Missing: 4 (2.5%)	No: 137 (87.3%) Yes: 19 (12.1%) Missing: 1 (0.6%)	Never: 148 (94.3%) Occasionally: 7 (4.5%) Usually/Always: 2 (1.3%)
Hallucinogens	No: 118 (75.2%) Yes: 35 (22.3%) Missing: 4 (2.5%)	No: 153 (97.5%) Yes: 0 (0%) Missing: 4 (2.5%)	Never: 157 (100%) Occasionally: 0 (0%) Usually/Always: 0 (0%)
Non-prescribed opioids	No: 135 (86.0%) Yes: 17 (10.8%) Missing: 5 (3.2%)	No: 151 (96.2%) Yes: 3 (1.9%) Missing: 3 (1.9%)	Never: 154 (98.1%) Occasionally: 1 (0.6%) Usually/Always: 2 (1.3%)

Main effects

ACEs demonstrated a significant positive relationship with CSI scores (standardized β = 0.434, P < 0.001, r² = 0.25), degree of opioid dependence symptoms (standardized β = 0.199, P = 0.021, r² = 0.043), lifetime polysubstance use (standardized β = 0.367, P < 0.001, r² = 0.109), depressive symptoms (standardized β = 0.262, P < 0.001, r² = 0.102), and anxiety symptoms (standardized β = 0.251, P = 0.001, r² = 0.095). ACEs were not significantly associated with MEDD (standardized β = -0.088, P = 0.43, r² = 0.003), pain severity (standardized β = 0.058, P = 0.44, r² = 0.008), current polysubstance use (standardized β = 0.200, P = 0.041 [not significant after FDR adjustment], r² = 0.047), pain-related activity interference (standardized β = 0.105, P = 0.13, r² = 0.037), or degree of polysubstance use to manage pain (standardized β = 0.221, P = 0.023 [not significant after FDR adjustment), r² = 0.049).

Above and beyond the effects of ACEs and study covariates, CSI scores were significantly associated with depressive symptoms (standardized β = 0.765, P < 0.001, r² = 0.543), anxiety symptoms (standardized β = 0.690, P < 0.001, r² = 0.472), and pain-related activity interference (standardized β = 0.491, P < 0.001, r² = 0.266). CSI scores did not show significant associations with MEDD (standardized β = 0.165, P = 0.074, r² = 0.008), degree of opioid dependence (standardized β = 0.078, P = 0.054, r² = 0.028), lifetime polysubstance use (standardized β = 0.168, P = 0.074, r² = 0.042), current polysubstance use (standardized β = 0.197, P = 0.057, r² = 0.076), polysubstance use to manage pain (standardized β = 0.205, P = 0.080, r² = 0.057) or pain severity (standardized β = 0.211, P = 0.027 [not significant after FDR adjustment], r² = 0.064).

Mediation models

See Figures 1–5. CSI symptoms significantly mediated the association between ACEs and depressive symptoms (LCL = 0.227, UCL = 0.438), between ACEs and anxiety symptoms (LCL = 0.212, UCL = 0.387), and between ACEs and pain-related activity interference (LCL = 0.124, UCL = 0.303). CSI scores did not mediate the association between ACEs and lifetime polysubstance use (LCL = -0.001, UCL = 0.070) or the association between ACEs and opioid dependence symptoms (LCL = -0.026, UCL = 0.108).

Figure 1.

Direct effects and mediation model for relationship between ACEs and anxiety, mediated by central sensitization symptoms.

Note: All coefficients are presented in standardized form.

Note: * P < .05; ** = P < .001.

Note: All path estimates include covariates for age and gender.

Figure 2.

Direct effects and mediation model for relationship between ACEs and depressive symptoms, mediated by central sensitization symptoms.

Note: All coefficients are presented in standardized form.

Note: * P < .05; ** = P < .001.

Note: All path estimates include covariates for age and gender.

Figure 3.

Direct effects and mediation model for relationship between ACEs and pain-related activity interference, mediated by central sensitization symptoms.

Note: All coefficients are presented in standardized form.

Note: * P < .05; ** = P < .001.

Note: All path estimates include covariates for age and gender.

Figure 4.

Direct effects and mediation model for relationship between ACEs and history of polysubstance use, mediated by central sensitization symptoms.

Note: All coefficients are presented in standardized form.

Note: * P < .05; ** = P < .001.

Note: All path estimates include covariates for age and gender.

Figure 5.

Direct effects and mediation model for relationship between ACEs and opioid dependence symptoms, mediated by central sensitization symptoms.

Note: All coefficients are presented in standardized form.

Note: * P < .05; ** = P < .001.

Note: All path estimates include covariates for age and gender.

Discussion

We aimed to investigate the relationships among childhood trauma, central sensitization, and substance, mood, and pain-related outcomes in a sample of adults on LTOT for a variety of CP conditions. Our hypotheses were partially supported. ACEs demonstrated a significant and moderately sized relationship with central sensitization, as well as significant but more modest positive associations with opioid dependence, lifetime polysubstance use, depressive symptoms, and anxiety symptoms. In contrast, ACEs did not show a significant association with pain severity or interference, current polysubstance use, MEDD, or degree of polysubstance use to manage pain. These findings suggest that childhood trauma is a potential risk factor for some substance use and mood-related outcomes in patients on LTOT; however, the lack of consistent effects suggests that there are likely other factors affecting these outcomes more robustly among people with CP. In our exploration of central sensitization as a potential mechanistic variable, we found that central sensitization was significantly associated with depressive symptoms, anxiety symptoms, and pain-related activity interference, but not other outcomes. In our exploration of central sensitization as a potential mechanistic variable, we found that central sensitization was significantly associated with depressive symptoms, anxiety symptoms, and pain-related activity interference, but not other outcomes. Consistent with our hypothesis, there was a stronger association between central sensitization symptoms and some key outcomes (mood disturbance and pain interference) than was noted for the associations between ACEs and these variables, suggesting that central sensitization may indeed be a more proximal indicator to pain-related function. This conclusion was further supported by our mediation analysis: Central sensitization mediated the association between ACEs and depressive symptoms, between ACEs and anxiety symptoms, and between ACEs and pain-related activity interference. However, no relationship was noted between ACEs and any substance-related outcomes, contrary to our hypothesis.

Overall, these findings highlight the complex nature of the relationship between ACEs and later outcomes. Results support prior research indicating that ACEs are a risk factor for depression, anxiety, and some opioid and substance-related outcomes among people with CP, and specifically individuals on opioids. However, there was not a significant direct relationship between ACEs and pain-related outcomes, including pain severity and interference, which have been found in prior studies (e.g.,^49,50). The mediation models indicated that ACEs did predict pain interference when central sensitization is included as a mediator, indicating that childhood trauma appears to be a particularly notable risk factor for more pain-related functional impairment as well as depression and anxiety when central sensitization is higher. From a psychoneuroimmunological perspective, childhood trauma exposure can result in altered stress reactivity and physiological sensitization to stress⁵¹ for some individuals, who may be at higher risk for poorer pain and mood-related outcomes. These findings support the growing understanding that an individual’s pain experience is determined not only by nociception and peripheral damage, but also by complex, evolutionary drives to predict threat and avoid harm.⁵²

Our substance use-related hypotheses were mostly unsupported. While ACEs were positively associated with severity of opioid dependence symptoms and the likelihood of lifetime polysubstance use, these relationships were not mediated by central sensitization, nor were ACEs predictive of current polysubstance use, MEDD, or using substances to cope with pain. Notably, substance use was fairly limited overall in this community sample, so findings may differ in higher-risk or substance use disorder (SUD)/CP populations. Our instrumentation for capturing these domains of substance use was also fairly low-resolution (endorsement of use of substances, all treated equally, using a yes/no format), which may have suppressed or otherwise affected the degree of available variance within the sample. Future research may benefit from other approaches to quantifying polysubstance use, such as daily frequency or dosage of substances used⁵³ or use of weighting or latent class analysis to establish empirically-derived classes of polysubstance use.⁵⁴ It should also be noted that the RODS instrument assesses indices of opioid dependence and misuse akin to OUD and does not capture other potential reasons for opioid use, such as pursuit of pain relief. It may be worthwhile for future research to examine whether ACEs and central sensitization symptoms are meaningfully associated with variability in risk-related patterns of opioid use for strictly analgesic purposes.

It is important to highlight that this sample reported a relatively high rate of ACEs overall (average 3.4, with 41% of the sample reporting 4 or more ACEs), as well as fairly high pain and disability levels, which is consistent with prior research on individuals on LTOT.^13,55 These sample characteristics may have contributed to the modest relationships among ACEs and pain outcomes. Still, more broadly, ACEs are generally a limited predictor of immediate outcomes, in part due to their nature as a distal variable that doesn’t change over time unlike the outcome variables, which supports exploration of more proximal indicators such as central sensitization that may represent downstream changes associated with ACEs. Additionally, the predictive power of ACEs is limited by the fact that many people recover from or overcome the effects of trauma/ACEs over time. Recent research suggests it is also important to capture and model positive social roles and support in childhood to better understand the impact of ACEs on chronic pain,⁵⁶ so future research models may show better predictive value if they incorporate both risk factors (like central sensitization) and markers of resilience against the deleterious effects of ACEs.

Additional limitations include the relatively small sample size and largely female and Caucasian demographic. This demographic makeup is not inconsistent with data on gender and racial disparities in chronic pain prevalence and treatment, which indicate that women have higher rates of CP,⁵⁷ women are more likely to be prescribed opioids for pain,⁵⁸ and White individuals receive more opioid prescriptions that racially minoritized individuals.⁵⁹ Generalizability to other demographic groups or chronic pain samples, or those not on LTOT, is unknown, and results should be replicated with larger and more diverse samples It would be useful to explore these relationships amongst individuals with trauma exposure other than (or in addition to) ACEs, as well as those with comorbid PTSD and substance use disorder. The ResearchMatch.org recruitment source should also be noted as a potential limitation. This large pool of volunteers is comprised of individuals interested in engaging in research, which provides strengths (e.g., attentiveness and little missing data), but it is possible they are different in some ways from those who do not choose to participate in research. As an example, the sample was generally highly educated, although a sizeable portion were not working or disabled. Unfortunately, race but not ethnicity was assessed due to a data collection error, so the complete demographic makeup is not ascertainable.

Conclusions

In a sample of adults with a variety of CP conditions on LTOT with relatively high childhood trauma, pain, and disability, ACEs were significantly associated with multiple mood and opioid risk-related outcomes. Further, central sensitization was identified as a potential mechanistic variable to help explain the relationship between ACEs and pain interference, depression, and anxiety. These results suggest that the Central Sensitization Inventory may serve as a useful clinical tool to identify elevated risk for mood disturbances and pain-related functional impairment among people with CP conditions who take opioids. Given its brevity and psychometric validity, the CSI could be integrated into routine assessments to flag individuals who may benefit from multimodal pain management or psychological support. Further research is needed to examine these relationships among larger, more diverse samples, as the elucidation of potential mechanisms of risk among individuals using LTOT could inform tailored prevention and treatment approaches to improve safety and overall outcomes.

Authors Contributions

All listed authors meet authorship criteria.

Funding

This work was supported by the National Institutes of Health/National Institute on Drug Abuse, K23DA058785-02, Taylor B. Crouch (PI)

Conflicts of interest

None declared.

Contents do not represent the views of the Department of Veterans Affairs or the United States Government.