Childhood Trauma Predicts Cancer Treatment–Related Pain in Breast Cancer Survivors

Abstract

Background:

Childhood trauma has been linked to neuropathic pain in noncancer populations, but its relationship with cancer treatment-related neuropathic pain is unknown.

Obiective

This secondary data analysis of a prospective, longitudinal, observational study aimed to explore the relationship of childhood trauma experience with pain severity, pain interference, and neuropathic symptom severity (NSS) 12 months after surgery in women receiving treatment for stage 0 to III breast cancer.

Methods:

Women (N = 44) recruited from a comprehensive cancer center self-reported childhood trauma experience, pain severity, pain interference, NSS, co-occurring symptoms, and pain beliefs via questionnaires. Descriptive statistics were used to describe childhood trauma experience. Linear regression was used to model childhood trauma and other predictors on pain variables 12 months after surgery.

Results:

Childhood trauma predicted pain severity and pain interference 12 months after surgery (P <.05), as did baseline pain severities and helplessness-pain catastrophizing. Age predicted only NSS. Together, the best models predicted 31.6% to 40.9% of the variance in pain severities at 12 months (P <.001).

Conclusions:

Childhood trauma exposure was a significant predictor of pain 12 months after breast cancer surgery and adjuvant treatment. Younger and helplessness-pain catastrophizing women are also at risk. Research is needed to identify preventive neuropathic pain interventions for high-risk women.

Implications for Practice:

Women receiving breast cancer treatment should proactively be assessed for childhood trauma history, possibly by using discreet previsit questionnaires. Childhood trauma survivors may be at high risk for poor pain outcomes and may benefit from tailored pain interventions.

Background

Pain is among the most common and feared treatment-related symptoms reported by individuals with cancer.¹ Specifically, among breast cancer survivors, 1 in 4 women has moderate to severe neuropathic pain after breast cancer treatment, which may include surgery, chemotherapy, radiation, endocrine, or immunotherapy.^2–4 Furthermore, breast cancer survivors who have experienced childhood trauma may be at even higher risk of developing severe treatment-resistant neuropathic pain. In the general population, about half of adults who report previous childhood trauma—physical, sexual, verbal, or other adverse psychological experiences before 18 years of age—also have chronic pain.⁵

Chronic neuropathic pain (from peripheral nerve damage) and nociceptive pain (from nonneuron tissue injury) may induce central nervous system (CNS) changes, leading to centralized pain: neuropathic pain that originates from the impairment of CNS pain processing mechanisms and persists even after the peripheral tissue has apparently healed. Centralized pain often co-occurs with other symptoms and can significantly interfere with quality of life.^6,7

Childhood trauma, among other factors such as chronic wide spread pain, fatigue, sleep disturbance,⁸ anxiety and depression,^3,8 and pain hypersensitivity,⁹ may be associated with poorly managed centralized pain. Further, maladaptive cognitions (ie, pain catastrophizing,^10–12 a low degree of belief in internal pain control¹³) may interact with these factors through central neuronal pathways and lead to more severe and persistent centralized pain. Specifically, previous evidence suggests a neurobiological link between childhood trauma and centralized pain^5,7,14,15 that could be mediated by changes in the amygdala and thalamus: emotion-and pain-processing structures of the brain.¹⁶

Although abundant research in noncancer populations supports the link between childhood trauma and centralized pain,^11,17 we found no literature that evaluates this relationship in the context of cancer treatment–related centralized pain. Although the broad relationship between childhood trauma and centralized pain may be generalizable across pain conditions, understanding the subtleties of this relationship in individual pain conditions is crucial for guiding appropriate assessment and tailored interventions. Specific types of childhood trauma may distinctly influence centralized pain development, co-symptomology, and pain treatment response in women receiving breast cancer surgery and several months of adjuvant treatment compared to populations with other types of centralized pain.

Specific Aims and Hypotheses

The primary aim of this prospective, longitudinal, observational study was to explore the relationship of childhood trauma experience with pain severity, pain interference, and neuropathic symptom severity 12 months after surgery in women receiving nonmetastatic breast cancer treatment. We hypothesized that childhood trauma would significantly predict higher pain severity, pain interference, and neuropathic symptom severity 12 months after surgery in this population. The 12-month time point was chosen to best represent chronic centralized pain outcomes after women had received all adjuvant treatments and started endocrine therapy.

Conceptual Model

Figure 1 provides a depiction of the study’s guiding conceptual model, partially derived from the Theory of Unpleasant Symptoms.¹⁸ According to the Theory of Unpleasant Symptoms, bi-directionally related physiological, psychological, and situational influencing factors may interact and influence the symptom experience. The adapted model illustrates the bidirectional relationships among pain characteristics and physiological, psychological, and situational factors that may accelerate or worsen the development of centralized pain after surgery and other cancer treatment–induced nerve injury. Specifically, preexisting and acute postsurgical pain,^19,20 widespread pain,¹⁹ younger age, female gender,^3,19,20 and genetic factors^21,22 are physiological factors that may predispose cancer survivors to develop centralized pain. Known psychological influencing factors of centralized pain include catastrophizing¹² and disempowered beliefs about pain control.^23,24 Childhood trauma, the key independent variable in our study, may be a key situational influencing factor of centralized pain.^25–28 Individuals who have risk factors, such as childhood trauma. and subsequently experience tissue or nerve injury from treatment (eg, axillary lymph node dissection)^3,19,20 may develop markedly more intense centralized pain and co-occurring symptoms (eg, fatigue and sleep disturbance).⁸

Figure 1.

Conceptual model of childhood trauma and cancer treatment–related neuropathic pain.

Methods

A secondary data analysis was conducted of a prospective, longitudinal, observational study with convenience sampling. Women (N = 44) with newly diagnosed stage 0 to III breast cancer scheduled to undergo breast surgery (mastectomy or lumpectomy) were recruited at a National Cancer Institute–designated Comprehensive Cancer Center between August 2012 and August 2014. Participants were recruited through brochures and presentations at breast cancer treatment workshops for individuals newly diagnosed with breast cancer. The study was also posted on the University of Michigan Clinical Studies and Chronic Pain and Fatigue Research Center websites. The study was approved by and conducted with oversight from the University of Michigan IRBMED and all participants provided written informed consent before study enrollment.

Eligibility Criteria

In addition, women were included if they were (1) 25 years or older and (2) able to read and speak English. They were excluded if they had (1) received neoadjuvant treatment; (2) preexisting carpal/tarsal tunnel syndrome, diabetic neuropathy, or arthritis in the hands or feet, which would confound neuropathic symptom measurement in the extremities; (3) psychiatric illness potentially impeding the individual’s ability to complete the study; or (4) life expectancy less than 1 year per physician report (due to the longitudinal nature of the study). Pregnancy was an additional exclusion criterion. Participants were permitted to take concomitant analgesic, psychotropic, and neuroleptic medications (except for investigational drugs) during the study.

Measures

Childhood trauma experience (independent variable) was self-reported before surgery (baseline). Via questionnaires at baseline and 12 months after surgery, participants self-reported pain severity, pain interference, and neuropathic symptom severity (dependent variables) and fatigue, sleep disturbance, sleep-related impairment, widespread pain, anxiety, pain catastrophizing, and beliefs in pain control (covariates). Patients reported demographic characteristics (age, marital status, education, and race) at baseline; cancer treatment data were abstracted from the medical record at 12 months.

THE CHILDHOOD TRAUMATIC EVENTS SCALE

The Childhood Traumatic Events Scale (CTES)²⁹ was used to measure childhood trauma experience. Participants self-reported experience of various childhood trauma types before 17 years of age: death of a family member or friend, parental upheaval, illness, violent abuse, sexual abuse, or other types of trauma. Per the CTES scoring manual, each individual type of trauma was coded 0 (no trauma), 1 (trauma with confiding [in others]), or 2 (trauma without confiding). An overall ordinal trauma experience variable was also coded: 0 (no trauma), 1 (trauma–confiding), and 2 (trauma–nonconfiding). Individuals who had confided in others for all types of trauma experienced were coded 1 (trauma–confiding) and termed “confiders.” Survivors who had not confided in others about any types of trauma they experienced were coded 2 (trauma–nonconfiding) and termed “nonconfiders.” Nonconfiders were considered at highest risk for adverse pain and health outcomes, based on the literature suggesting that lack of confiding could be associated with greater morbidity.^17,30 The CTES has been widely used and its validity has been supported by psychosocial experimentation²⁹; associations with cortisollevels³¹; and self-reported chronic pain,¹¹ emotional regulation, psychological distress, anxiety, poorer mood, and symptom burden.³² Each item’s content validity index was 0.75 or greater, based on established content validity evaluation methods³³ and reviews from 2 PhD-prepared nurses, a pain and health psychologist who has used the CTES, and a pain clinic social worker (unpublished).

BRIEF PAIN INVENTORY SHORT FORM

Pain severity and pain interference were evaluated using the Brief Pain Inventory Short Form, which measured total pain severity (4 items) and pain interference (7 items). Each item was rated on a numeric rating scale ranging from 0 (no pain; does not interfere) to 10 (pain as bad as you can imagine; completely interferes). Total pain severity was obtained by averaging the participant’s “least,” “worst,” “average,” and “now” pain items. Pain interference was an average of the items measuring the degree pain interfered in the last 24 hours with mood, enjoyment of life, and 5 daily activities. Higher scores indicated worse pain severity and interference. Evidence supports strong internal consistency, stability, and validity of the Brief Pain Inventory Short Form in cancer populations.^34,35

PainDETECT

The 9-item PainDETECT was used to measure neuropathic symptom severity. The sum of the items, on a scale from −1 to 38, indicated the likelihood that the pain was neuropathic,³⁶ termed “centralized pain severity” in our study. Seven items measured the degree of neuropathic pain qualities experienced (eg. numbness, tingling, burning, radiating) on a 0 (never) to 5 (very strongly) Likert scale. One item (coded 0, −1, or +1) asked participants to select the best depiction of their pain pattern. The last item asked about the presence of radiating pain (0, yes; 2, no). A sum score of 12 or lower indicates likely nociceptive pain; 19 or higher indicates likely neuropathic pain. Evidence supports the instrument’s construct validity, stability, sensitivity, and specificity.³⁶

COVARIATES

Abundant evidence supports the validity and reliability of the following symptom and pain cognition measures: the Patient-Reported Outcomes Measurement Information System (PRO:MJS) instruments,^37,38 Michigan Body Map,^39,40 State-Trait Anxiety Inventory Y-2Ax form,^39,40 Pain Catastrophizing Scale (PCS),^41,42 and Beliefs in Pain Control Questionnaire.^43,44

Patient-Reported Outcomes Measurement Information System:

The Patient-Reported Outcomes Measurement Information System short forms were used to measure symptoms of fatigue (7 items), sleep disturbance (8 items), and sleep-related impairment (8 items). Each PROMIS item was rated on a 1 (never/not at all) to 5 (always/very much) Likert scale; the total raw scores, ranging 5 to 35 (fatigue) and 8 to 40 (sleep disturbance and sleep-related impairment), were transformed linearly to T scores. Higher scores indicated worse symptoms.

Michigan Body Map:

The Michigan Body Map measured the number of self-reported chronically (present for at least 3 months) painful sites selected from 35 labeled body sites on an illustration of a person in the anatomical position. Higher scores indicated more widespread pain.

State-Trait Anxiety Inventory Y-2 Ax form:

The 10-item State-Trait Anxiety Inventory Y-2 Ax form measured anxiety on a scale from 10 to 40. Individual items were rated on a 1 (almost never) to 4 (almost always) Likert scale. Higher scores indicated higher anxiety.

Pain Catastrophizing Scale:

The Pain Catastrophizing Scale was used to measure an individual’s evaluation of pain. Each of the 13 items was rated on a 0 (not at all) to 4 (all the time) Likert scale. The items were divided into 3 subscales: rumination (scored 0–16), magnification (0–12), and helplessness (0–24). A total score was also calculated from the sum of the 13 items (scored 0–52). Higher scores indicated greater pain catastrophizing.

Beliefs in Pain Control Questionnaire:

In 3 summed subscales, the Beliefs in Pain Control Questionnaire measured an individual’s perceived ability to control pain across 3 factors: one’s own actions (“internal locus of pain control” 5-item scale scored 5–30) or external medical treatment and chance happening (“external locus of pain control” from “powerful doctors” and “chance happening,” each 4item scale scored 4–24). Higher scores on any of the scales indicated stronger beliefs in internal or external loci of pain control

Data Analysis

All statistical analyses were run using R version 3.3.1. Given the exploratory nature of this research, no power analysis was conducted for this study. Each variable was described using means, standard deviations (SDs), and frequencies. Chi-square and independent t tests were used to evaluate the differences between individuals who had and who had not experienced childhood trauma. In addition, density plots were used to evaluate the distributions of the 3 outcome variables: pain severity, pain interference, and neuropathic symptom severity. A 2-sided P value ≤.05 was considered statistically significant.

Missing cases were checked for systematic causes manually and with automatic analytic techniques. Missing data for the primary aim were handled using multivariate imputation by chained equations. All individuals (N = 44) were included in the descriptive analysis; however, 2 individuals were excluded from the inferential statistical analyses. One individual was excluded because he/she did not complete the CTES. The 1 confider—childhood trauma survivor who had confided in others for all types of trauma experienced—was excluded from the analysis, because evidence suggests that health outcomes of trauma survivors who have social support differ from those who lack social support³⁰ and from childhood trauma-free individuals.¹⁷

Analysis of the Primary Aim

The primary aim was to explore the relationship of childhood trauma experience with pain severity, pain interference, and neuropathic symptom severity 12 months after surgery. The baseline and 12-month differences in participant characteristics between individuals who had and had not experienced childhood trauma (excluding the confiding individual) were evaluated with independent t tests. The between-group differences in baseline to 12-month change scores for the pain outcomes were also evaluated with independent t tests. Bivariate Spearman correlations and analysis of variance were used to evaluate associations of the covariates with the pain outcomes and with childhood trauma. Covariates were added to the predictive models if they were significantly associated with the pain outcomes (P ≤.05) but not with childhood trauma experience. Linear regression was used to model child hood trauma and the respective baseline pain severities on pain severity, pain interference, and neuropathic symptom severity at the 12-month time point in separate models. Additional predictors were tested in the models and analysis of variance was used to compare the models.

Results

Although 63 patients were enrolled, 7 participants were later deemed ineligible due to diagnosis of recurrent or stage IV cancer (n = 3), arthritis (n = 2), pregnancy (n = 1), or scheduled surgery outside of the study site (n = 1). In addition, 12 participants withdrew before baseline measurements for various reasons: time constraints/scheduling complications (n = 4), lass to follow-up (n = 6), and unknown reasons (n = 2). Of the participants who completed baseline assessments and were included in the analysis (n = 44), 42 (95.5%) completed 12-month assessments. Each questionnaire item included in the analysis was answered by at least 75% of the participants.

Sample Characteristics

Table 1 describes the baseline (presurgical) characteristics of the sample by childhood trauma experience. Briefly, the sample (N = 44) consisted of women with a mean (SD) age of 56.52 (11.25) years (range, 31–82 years), most of whom were white (90%) and married (61%) and had a baccalaureate degree or higher (59%). Most of the women had stage I (50%) or stage II (39%) breast cancer. Of the 59% who had undergone lumpectomy, 3 also underwent a mastectomy. Six (14%) had axillary lymph node dissection. and 12 (27%) had immediate reconstruction. On average, the women had undergone 2.05 (SD, 1.23) breast surgical procedures, and most had received some form of adjuvant treatment.

Table 1 •.

Demographic and Clinical Characteristics by Childhood Trauma Experience (N = 44)

Characteristic	Total (N = 44)	Childhood Trauma Experience		P
		No Trauma (n = 21)	Trauma—Nonconfiding (n = 21)
Age, mean (SD), range	56.52 (11.3), 31–82	58.38 (11.8), 31–82 n (% of total)	54.81 (10.32), 36–73	.303
Marital status				.196
Married	27 (61)	14 (67)	12(57)
Separated or divorced	5(11)	0	5(24)
Other	12(27)	7(33)	4(19)
Education (n = 43)				.773
Associate degree or less	17 (40)	6(29)	10 (48)
Bachelor degree or Higher	26 (59)	15 (71)	11 (52)
Race				.551
White	40 (90)	20 (95)	18 (86)
African American	2(5)	K5)	K5)
Other	2(5)	0	2(10)
Breast cancer stage				.128
Stage 0	2(4)	0	2(10)
Stage I	22 (50)	9(43)	12(57)
Stage II	17 (39)	9(43)	7(33)
Stage III	3(7)	3(14)	0
Surgerya
Mastectomy	21 (48)	9(43)	12(57)	.537
Lumpectomy	26 (59)	14 (67)	10 (48)	.35
Axillary lymph node dissection	6(14)	5(24)	K5)	.186
Immediate reconstruction	12(27)	4(19)	8 (38)	.306
Adjuvant treatmenta
Chemotherapy	16 (36)	10 (48)	5(24)	.198
Radiation therapy	28 (64)	16(76)	10 (48)	.112
Endocrine therapy	30(77)	14 (67) Mean (SD), range	14(67)	1
Self-reported
Pain severity	1.79(2.58), 0–11.33	1.17 (1.63), 0–5.67	2.57 (3.2), 0–11.33	.085
Pain interference	1.16 (1.98), 0–8.14	0.77 (1.24), 0–3.71	1.71 (2.57), 0–8.14	.16
Neuropathic symptom severity	8.78 (8.39), 1–32	4.5 (3.87), 1–11	11.29 (9.37), 1–32	.015
Fatigue	50.05 (7.81)	48.24 (6.64)	52.69 (7.85)	.057
Sleep disturbance	50.91 (10.53)	45.66 (9.75)	56.7 (8.2)	.000
Sleep-related Impairment	48.07(11.38)	44.62 (9.99)	52.2 (11.46)	.028
Widespread pain	2.33 (4.4)	2.06 (3.19)	2.75 (5.52)	.67
Anxiety	15.98 (4.45)	15.68 (4.73)	16.45 (4.41)	.604
Pain catastrophizing	8.93 (8.88)	6.9 (8.89)	11.14 (8.83)	.129
Helplessness	3.34 (3.96)	2.29 (3.39)	4.62 (4.33)	.059
Magnification	1.91 (1.79)	1.57 (1.8)	2.33 (1.77)	.175
Rumination	3.68 (4.21)	3.05 (4.42)	4.19 (3.89)	.379
Beliefs in pain control
Internal	16.55 (4.16)	16.1 (4.64)	16.71 (3.41)	.625
External	9.32 (3.23)	9.05 (3.37)	9.33 (3.06)	.775
Powerful doctors	10.77 (3.59)	10.62 (2.97)	10.48 (4.04)	.897

The sole individual who had experienced trauma but confided in others was excluded from the t test analysis. Another individual did not complete the childhood trauma survey.

^aSome individuals had a combination of treatments (eg, women could have received both a lumpectomy and mastectomy).

On average, women reported moderate severities of fatigue (mean [SD], 50.05 [7.81]), sleep disturbance (mean [SD], 50.91 [10.53]), sleep-related impairment (mean [SD], 48.07 [11.38]), and anxiety (mean [SD], 21.71 [2.9]) before surgery. Women had a mean (SD) of 2.33 (4.4) (range, 0–20) chronically painful sites and displayed low pain catastrophizing (mean [SD], 8.93 [8.88]) and a moderate level of belief in internal locus of pain control (mean [SD], 16.55 [4.16]) before surgery.

Pain Characteristics

Figure 2 portrays the changes in participants’ (a) mean and (b) median pain severity, pain interference, and neuropathic symptom severity from baseline to 12 months in individuals who had and had not experienced childhood trauma. For the entire sample at baseline, mean (SD) pain severity was 1.79 (2.58), pain interference was 1.16 (1.98), and neuropathic symptom severity was 8.78 (8.39). Four (9%) participants at baseline and 1 of these 4 individuals at 12 months—all of whom had experienced childhood trauma—had likely neuropathic pain (PainDETECT score ≥19). No individuals had developed new neuropathic pain at 12-month follow-up. The means showed no significant increase in pain severities over time.

Figure 2.

Pain severities at baseline and 12 months by childhood trauma experience.

Childhood Trauma Experience

Figure 3 displays then (%) of trauma-free women, confiders, and nonconfiders for each trauma type. The most common types of trauma experienced were “other” self-reported traumas (n = 13; 30.2%), “parental upheaval” (n = 8; 18.6%), and “death of a family member or friend” (n = 7; 16.3%). No individuals had previously confided in others about traumas related to parental upheaval, violence, or sexual abuse. Thirteen individuals (1 confider and 12 nonconfiders) had experienced multiple types of traumas. Overall, 21 women had experienced childhood trauma without confiding.

Figure 3.

Childhood trauma experience by individual trauma type.

Neither demographic nor clinical variables differed significantly between trauma-free and nonconfiding individuals. However, neuropathic symptom severity (P =.015; confidence interval [CI], −12.13 to −1.46), sleep disturbance (P<.001; CI, −16.67 to −5.42), and sleep-related impairment (P =.028; CI, −14.3 to −0.88) were significantly higher in nonconfiders than trauma-free women. Nearing significance, mean fatigue (P =.057; CI, −9.04 to 0.14) and helplessness-pain catastrophizing (P =.059; CI, −4.76 to 0.1) were also higher in the nonconfiders.

Childhood Trauma Experience Predicts Pain at 12 Months

Table 2 summarizes the regression models conducted for each pain outcome. Basic (no additional covariates) and complex: (with covariates) models are shown for each outcome, along with an F statistic comparing the models. When compared with nontraumatized women, individuals who had experienced childhood trauma reported significantly higher pain severity (CI, −1.76 to −0.66), pain interference (CI, −2.03 to −0.73), and neuropathic symptom severity (CI, −3.84 to −1.02) at 12 months (all P ≤.001). Furthermore, the percentage of individuals at risk for or with likely neuropathic pain was significantly higher in individuals who had experienced childhood trauma (x² = 15.82, p <.001).

Table 2 •.

Regression Models of Childhood Trauma and Other Predictors of Pain Outcomes (N = 44)

	β(SE)					Adi r2	F
Model	Intercept	Childhood Trauma	Baseline Pain	Helpless-Pain Catastrophizing	Age
Pain Severity							3.21
Aa	0.83 (0.17)a	0.25 (0.12)b	0.61 (0.06)a	-	-	0.365
Ba	0.75 (0.17)a	0.22 (0.12)c	0.55 (0.07)a	0.06 (0.03)c	_	0.371
Pain Interference							9.39a
Aa	0.64 (0.19)a	0.3 (0.13)b	0.74 (0.07)a	-	-	0.387
Ba	0.46 (0.2)b	0.22 (0.13)	0.61 (0.08)a	0.12 (0.04)a	-	0.409
Neuropathic Symptom Severity							23.34a
Aa	5.71 (0.5 l)a	0.29 (0.35)	0.29 (0.05)a	-	-	0.176
Ba	9.22 (1.62)a	-0.01 (0.32)	0.1 (0.05)b	0.61 (0.09)a	−0.07 (0.03)b	0.316

This table presents the statistical values for separate linear regression models for each pain outcome: pain severity, pain interference, and neuropathic symptom severity. Each outcome is modeled (a) on just childhood trauma and the respective baseline pain score alone, then (b) on additional covariates. The F statistic indicates the degree of improvement from model A to model B.

Abbreviations: A, basic model; B, complex model.

^aSignificant at P < .001.

^bSignificant at P < .05.

^cSignificant at P < .1.

In the 3 basic regression models, childhood trauma and the respective baseline pain severities predicted 27.5% of the variance in pain severity, 38.5% of the variance in pain interference, and 17.6% of the variance in neuropathic symptom severity at 12 months (all P ≤.005). Improvement in pain severity (F = 3.21; P =.074), pain interference (F = 9.69; P =.002}, and neuropathic symptom severity (F = 39.04; P <.001) models was found by adding helplessness-pain catastrophizing as a covariate. Together, the variables predicted 37.1% of the variance in pain severity, 40.9% of the variance in pain interference, and 29.9% of the variance in neuropathic symptom severity at 12 months. Ultimately, childhood trauma was not a significant predictor of neuropathic symptom severity. The model with baseline neuropathic symptom severity (β=.1, SE 0.05), helplessness-pain catastrophizing (β =.61, SE 0.09), and age (β = −.07, SE 0.03) was best without childhood trauma experience, predicting 32% of the variance in neuropathic symptom severity at 12 months.

Discussion

Our pilot study found that half the women in our sample had experienced some type of childhood trauma—most commonly from parental upheaval, death of a family member or friend, and other nonspecified events—without confiding in others. In addition, our pilot study provides preliminary evidence supporting the positive relationships between childhood trauma and both pain severity and pain interference in women 12 months after breast cancer surgery (who may have also received other adjuvant treatments). Higher helplessness-pain catastrophizing before surgery also predicted each pain outcome. Neuropathic symptom severity, at both baseline and 12 months, was significantly higher in nonconfiding trauma survivors than in women who did not experience trauma; however, childhood trauma did not predict neuropathic symptom severity. The best neuropathic symptom severity model showed that baseline neuropathic severity, helplessness-pain catastrophizing, and younger age predicted nearly one-third of the variance in neuropathic symptom severity at 12 months.

Previous literature suggests that about half of breast cancer survivors report childhood trauma.⁴⁵ Physical and emotional neglect may be most common, and sexual abuse the least common type of childhood trauma experienced,⁷ patterns that are consistent with the general adult population.⁵ The 2016 National Survey of Children’s Health, conducted by the Child and Adolescent Health Measurement Initiative, showed that at least 38% of children in every US state have experienced childhood trauma related to events such as death or incarceration of a parent, witnessing violence, and living with someone with a drug or alcohol problem⁴⁶

However, the true prevalence of childhood trauma experience in women receiving breast cancer treatment may exceed 50%. Our sample consisted primarily of well-educated white women, a population that may exhibit relatively low rates of childhood trauma. Although childhood trauma is prevalent among children of all races and socioeconomic statuses, Child and Adolescent Health Measurement Initiative reports higher rates of childhood trauma in Hispanic (51%) and black (64%) children, compared to Caucasian children (40%). Furthermore, childhood trauma has been estimated to affect 62% of children in low-income families versus 26% of children in high-income families.⁴⁶

Previous literature in noncancer populations has also linked childhood trauma experience with higher pain severity^25,28 and provides further insight into the nuances of the childhood trauma-pain relationship. For example, evidence suggests cumulative effects: A higher number of traumatic experiences is associated with higher chronic pain incidence.²⁸ Certain types of childhood trauma may be more closely associated with worse pain outcomes, including physical, sexual, and verbal abuse^15,25 and chronic fear of a family member.²⁸ In addition, evidence suggests that the period between 3 and 5 years of age may be most critical in predicting health outcomes related to childhood trauma experience.⁴⁶ Finally, supportive relationships can moderate the relationship between childhood trauma experience and health outcomes.⁴⁶

Evidence suggests common neurobiological pathways between childhood trauma and centralized pain, anxiety and mood disorders, and fatigue.^5,7,14,15 Anxiety and mood disorders¹⁵ and fatigue⁷ may partially mediate and moderate the relationship between childhood trauma experience and centralized pain. Furthermore, the link between childhood trauma (especially emotional and physical abuse)⁵ and cancer treatment–related centralized pain could be mediated by perception and report of pain; inferior coping,⁷ help-seeking, and pain treatment compliance and response^{10,12,23,24,27}; and reduced social support.⁵

Neurobiologically, the links among childhood trauma experience, centralized pain, and co-occurring symptoms may be due to childhood trauma-induced structural alterations in the amygdala and thalamus. The thalamus and amygdala process both emotional distress and peripheral pain signals and mediate the activation of other CNS structures responsible for neuronal, hormonal. cardiovascular, and behavioral changes.¹⁶ Cyclical reminders of previous traumatic experiences may lead to conditioned responses and breakdown of inhibitory mechanisms in the amygdala This long-term exposure to emotional distress and constant activation of the amygdala can lead to amygdalar changes, inflammation, and hypersensitivity of the pain pathways.⁴⁷

Most of the literature linking childhood trauma and pain is based on individuals with fibromyalgia and other spontaneously occurring chronic widespread pain conditions.^15,25,28 However, the mechanisms of centralized pain development after surgery and cancer treatment may differ from spontaneously occurring chronic pain conditions. Cancer treatment–associated psychological stress,⁴⁸ mechanically induced pain,⁴⁹ and co-occurring symptorns⁸ may interact uniquely with the neurobiological and psychological mechanisms of childhood trauma Furthermore, gender may influence the trauma-pain relationship. Specifically, women have exhibited stronger associations between childhood trauma and pain outcomes.¹⁵ By exploring the relationship of childhood trauma with pain outcomes 12 months after breast cancer surgery, this article contributes to the understanding of the specific relationship between childhood trauma and cancer treatment–related centralized pain in female cancer survivors.

Study Limitations

The findings of this study are exploratory and intended only to inform future research. First, our pilot study was not powered to detect all potentially significant centralized pain predictors and evaluate the nuances of the relationship between childhood trauma and centralized pain. Second, 1 participant confided in others about all traumatic experiences and was therefore excluded from the inferential statistical analyses. Third, only 4 participants at baseline and 1 of these 4 participants at 12 months exhibited likely neuropathic pain. Thus, we were unable to run logistic models to evaluate childhood trauma as a predictor of neuropathic pain incidence 12 months after breast cancer surgery. Furthermore, we were unable, because of small numbers, to evaluate mediators and moderators of the relationship of childhood trauma with the pain outcomes. Notably, we did not assess the influences of specific support types (eg, formal mental health services, family, and friends) or types of trauma on the relationship between childhood trauma experience and pain outcomes. Fourth, the study was conducted at 1 university National Cancer Institute–designated comprehensive cancer center with a well-educated and mostly white sample. Thus, the findings may not be generalizable to other women outside of the university environment. Finally, trauma-related confounding factors may have threatened the validity of the findings. For example, the study did not control for repeated traumatic experiences including trauma in adult-hood. posttrauma confiding behavior and receipt of mental health or other support services, and years since the traumatic experience.

Implications for Practice

In women receiving breast cancer treatment, clinicians should assess for childhood trauma experience to identify high-risk individuals before surgery who may need tailored pain interventions in addition to standard pain management. Proactive and discreet assessment of trauma history is particularly important to identify individuals who have not previously confided in or sought help from others after experiencing childhood trauma. Nonconfiding patients may be at higher risk for adverse pain outcomes throughout each survivorship phase of their cancer treatment than patients who have confided in others about their experiences of childhood trauma. Previsit medical record-linked electronic questionnaires may be ideal for capturing trauma history, especially among non–previously confiding individuals who may benefit most from early/preventative pain interventions. For example, supportive therapies (eg, cognitive behavioral therapy and mindfulness)¹⁰ or interventions specifically targeting unresolved emotional experiences in individuals with pain⁵⁰ may be important adjunctive treatments to commonly prescribed analgesic and psychotropic drugs for chronic centralized pain. Currently, opioids, gabapentin, pregabalin, tricyclic antidepressants, serotonin, norepinephrine reuptake inhibitors, anticonvulsants, ketamine, lidocaine, and similar drugs are often prescribed alone.¹ Our research highlights the complexity of pain mechanisms and individual moderators of centralized pain development, such as childhood trauma experience. Future research in this area may inform the tailoring of more effective pain interventions for women with breast cancer who have experienced childhood trauma.

Conclusions

About half of all women, including breast cancer survivors, report experiencing a traumatic event before 18 years of age. Childhood trauma may be a significant predictor of centralized pain severity in women before and 12 months after breast cancer surgery and potentially adjuvant treatment. However, future research is needed to confirm and evaluate the strength and potential mechanisms of the relationship between childhood trauma and cancer treatment–related centralized pain. Specifically, future studies should use larger and more diverse samples, more complex coding to represent the development of persistent centralized pain, and more comprehensive measurement of childhood trauma experience and important confounding factors. A greater understanding of the mechanisms and moderators of the childhood trauma-pain relationship is necessary to guide better centralized pain assessment, identification of high-risk women, and tailoring of interventions for women receiving breast cancer treatment.