Specific Associations Between Type of Childhood Abuse and Elevated C-Reactive Protein in Young Adult Psychiatric Rehabilitation Participants

Abstract

Background

Early-life adversity such as childhood emotional, physical, and sexual trauma is associated with later-life psychiatric and chronic medical conditions, including elevated inflammatory markers. Although previous research suggests a role for chronic inflammatory dysfunctions in several disease etiologies, specific associations between childhood trauma types and later-life inflammation and health status are poorly understood.

Methods

We studied patients (n = 280) admitted to a psychiatric rehabilitation center. Self-reported histories of childhood emotional, physical, and sexual trauma were collected with a standard instrument. At the time of admission, we also assessed individuals’ body mass index and collected blood samples used to examine inflammatory marker C-reactive protein (CRP) levels.

Results

The prevalence of all 3 types of abuse was relatively high at 21% or more. Fifty percent of the sample had elevations in CRP, with clinically significant elevations in 26%. We found that compared with a history of emotional or physical abuse, a history of childhood sexual trauma was more specifically associated with elevated CRP. This result held up when using linear regressions to examine the contribution of body mass index.

Limitation

Our sample was relatively young, with an average age of 27.2 years and minimal representation of ethnic and racial minority participants.

Conclusion

Relative to childhood emotional and physical trauma, childhood sexual trauma may lead to elevated inflammatory responses, as confirmed in our finding of an association between CRP and sexual abuse. Future studies need to assess the causal link between childhood sexual trauma and poorer health outcomes later in life.

Introduction

Childhood and adolescence are pivotal developmental periods during which physical and behavioral functions develop to shape later-life adaptive functioning. Thus, traumatic environmental experiences during critical periods for a developing individual can have profound and lasting negative consequences on developmental and later-life health outcomes (Schrepf et al., 2014; Kiecolt-Glaser et al., 2015; Petrov et al., 2016; Jonker et al., 2017; Moraes et al., 2017). For instance, previous research showed that childhood trauma (Gould et al., 2012), including sexual, emotional, and physical abuse or other forms of maltreatment, is associated with cognitive impairments and poses long-lasting negative health consequences and significant morbidity and premature mortality (Edwards et al., 2012; Lu et al., 2013; Matthews et al., 2014; Jabbi and Nemeroff, 2019; Pinto Pereira et al., 2019; Lippard and Nemeroff, 2020; Teicher et al., 2022).

Although research is beginning to identify the consequences of childhood adversity in terms of the negative impacts on the central nervous system and peripheral biomarkers, the precise linkage between childhood abuse history and later-life physical and psychological morbidity is not fully understood. Previous research revealed that over one-half of all developing children or their caregivers report at least 1 substantial traumatic experience before reaching 18 years (Saunders and Adam, 2014). Traumatic incidents of perceived or experienced violations of bodily integrity in the forms of emotional abuse, including the threat of physical harm, physical abuse, including actual injury resulting from violence, or the experience of sexual abuse can have far-reaching negative consequences in later-life emotional and physical well-being, morbidity, and poorer health outcomes (Scheffers et al., 2017; Lippard and Nemeroff, 2020). However, whether these different childhood trauma subtypes and the severity of their occurrences can differentially influence later-life mental and physical health outcomes is not well understood.

Early-life trauma increases the risk of several psychiatric disorders, including depression (Kiecolt-Glaser et al., 2015), anxiety disorders, and posttraumatic stress disorder (PTSD) (Choi and Sikkema, 2016; Gould et al., 2021). Moreover, even in domains of severe mental illness, early-life trauma appears to be associated with an increased risk for the development of schizophrenia among individuals equally predisposed by having a first-degree relative with the disease (Morgan and Anderson, 2016). Further, individuals with early-life trauma are more likely to develop treatment-resistant mood disorders and manifest multiple features of poor disease outcomes, including suicidal behaviors (Tunnard et al., 2014; Kiecolt-Glaser et al., 2015; Yrond et al., 2021).

More recently, the effects of early-life trauma on inflammatory responses and long-term dysregulations in immune functioning have been demonstrated (Pace et al., 2006; Kiecolt-Glaser et al., 2015; Danese and Lewis, 2017; Pitharouli et al., 2021). For example, circulating C-reactive protein (CRP), an acute-phase protein of hepatic origin that rises in response to inflammation, can be increased after early-life trauma (Kiecolt-Glaser, 2015; Pitharouli et al., 2021). Elevated levels of CRP are also associated with cardiovascular disease (Kiecolt-Glaser, 2015; Ananthan and Lyon, 2020; Pitharouli et al., 2021), obesity (Matthews et al., 2014), and certain cancers (Ananthan and Lyon, 2020). In addition, elevated CRP concentrations (Kiecolt-Glaser, 2015; Ananthan and Lyon, 2020; Pitharouli et al., 2021) in traumatized individuals later in life can be associated with not only abuse histories but also with elevated measures of body mass index (BMI), anxiety, and depressive symptoms (Noll et al., 2007).

The 3 types of early-life abuse measured with the Childhood Trauma Questionnaire (CTQ)—sexual, emotional, and physical abuse (Bernstein et al., 1994, 1997, 2003)—may have different implications for later-life physical and mental health outcomes. For instance, in a study of newly admitted adolescent inpatients, a history of emotional abuse, but not physical or sexual abuse, correlated with being cyberbullied immediately before admission to an inpatient psychiatric unit (Saltz et al., 2020). Likewise, in a study on the risk of developing PTSD in the immediate aftermath of trauma (Gould et al., 2021), the total severity of early-life trauma, as measured by the total CTQ score, was associated with risk for the development of PTSD following this additional traumatic experience. In addition, many studies have suggested that sexual abuse is particularly pernicious in risk for later-life psychopathology (Kiecolt-Glaser, 2015; Baumeister et al., 2016; Ananthan and Lyon, 2020; Pitharouli et al., 2021).

Matthews and colleagues examined the relationship between CRP levels and childhood sexual, emotional, and physical abuse and neglect in a cohort of mid-life women (Matthews et al., 2014). They found that sexual and emotional abuse history, history of emotional and physical neglect, and the total number of recorded incidents of childhood abuse were associated with elevated CRP levels over 7 years. However, these findings of correlations between childhood abuse history and the number of incidents with later-life CRP levels were moderated by BMI (Matthews et al., 2014). All participants were female, and all were outpatients. In addition, recent integrative reviews and meta-analyses examining the relationship between the history of childhood adversity and later-life measures of CRP, including measures of interleukin family of inflammatory factors and tumor necrosis factor-α, have found differential relationships between the type of trauma and later-life inflammatory measures, but the results of these meta-analyses were not clear-cut (Schrepf et al., 2014; Baumeister et al., 2016; Pinto Pereira et al., 2019; Brown et al., 2021).

In the present study, we examined young (mean age = 27 years) females and males, varying in their lifetime trauma history, at the time of admission to a psychiatric rehabilitation facility. We assessed the presence and severity of early-life trauma by documenting the history of (1) sexual abuse, (2) emotional abuse, and (3) physical abuse using the CTQ (Bernstein et al. 1994, 1997, 2003). Our study design affords a potentially reliable assessment of childhood trauma by assessing a young adult sample with limited chronicity of psychiatric and medical conditions and more proximal experiences of childhood trauma. Obtaining BMI and CRP data at the time of admission during the first physical examination precluded treatment-related changes in CRP or BMI from influencing the results. Furthermore, these assessments at the time of admission allowed the exclusion of individuals with chronic medical conditions that could cause elevations in inflammatory markers. We specifically examined next the relationship between sexual, emotional, and physical abuse/trauma history (particularly meeting criteria for a definite history of abuse): (1) levels of CRP, and (2) BMI/obesity. Childhood sexual abuse has pernicious effects on later-life mental and physical health, including increased prevalence of later-life mood disorders and elevations in BMI (Moraes et al., 2017) and increased association between CRP levels and obesity in later-life depression (Pitharouli et al., 2021). Because elevated CRP and BMI are recognized as biophysical consequences of abuse (Noll et al., 2007; Matthews et al., 2014; Kiecolt-Glaser, 2015; Ananthan and Lyon, 2020; Pitharouli et al., 2021), we examined associations between BMI and CRP with CTQ scores of abuse history subtypes.

Methods

Participants

This study comprised patients admitted to Skyland Trail, a non-profit residential, partial hospital, and intensive outpatient psychiatric rehabilitation facility in Atlanta, GA, USA. The center has a continuum of care in the context of a recovery model. As symptomology improves, more symptomatic individuals who are initially placed in residential facilities transition into day treatment, intensive outpatient, and transitional housing treatment tracks. Therefore, less symptomatic individuals were directly entered into day treatment or intensive outpatient tracks. Individuals with chronic physical illnesses cannot be managed at Skyland Trail and are not admitted to residential services. Patients were included in the current analyses if they were accepted between April 2019 and December 2020, completed the CTQ, had a physical examination with a BMI score, and a blood sample collected at admission and assayed for CRP (n = 280).

Participants are not admitted to Skyland Trail if they have evidence of significant risk for suicidal or homicidal behavior. Thus, a very recent suicide attempt precludes admissions. Given potential confounding conditions such as chronic infection, taking antibiotics, 48 hours of non-steroidal anti-inflammatory drug (NSAID) medication use prior to the blood collection for CRP assay, etc., we excluded all individuals with these potential underlying inflammatory confounds that could influence CRP measures from this study. Diagnostic information was collected with the Mini-International Neuropsychiatric Inventory (MINI). Diagnoses using DSM-5 criteria were generated at admission with a consensus process involving trained clinicians who administered the structured rating scale and an attending psychiatrist who reviewed all available records and the clinician impressions. These diagnoses were collected at admission using a structured procedure on which we had previously published the methods (Kotwicki and Harvey, 2013).

The institutional review board (IRB) at the University of Miami Miller School of Medicine evaluated this project. That IRB determined that because all data were obtained for clinical purposes and were examined completely de-identified, the project is exempt from IRB review. See this determination in eMaterials of a prior publication (available at https://academic.oup.com/ptj).

Assessments

After admission and during their rehabilitation stay, patients at the center received biweekly clinical assessments tailored to their primary treatment targets. Participants with a diagnosis of major depression from the MINI interview were assessed with the Montgomery-Åsberg Depression Rating Scale and the Hamilton Anxiety Rating Scale. Participants with a MINI diagnosis of bipolar disorder were also evaluated with the Montgomery-Åsberg Depression Rating Scale and the Young Mania Rating Scale. In addition, the Brief Psychiatric Rating Scale (BPRS) was used to assess all cases with psychotic symptoms, including participants with major depression or bipolar disorder with psychotic features. Ratings were performed by raters who were not the primary clinicians. Thus, not all participants were rated with the same clinical assessment scales, for which we created an aggregate for use in statistical analyses as described below.

Childhood Trauma Questionnaire

—The CTQ is a 28-item self-report questionnaire including questions about childhood abuse or neglect experiences (Bernstein et al., 1994). It consists of 5 subscales measuring childhood sexual, emotional, or physical abuse and emotional or physical neglect (Bernstein et al. 1994, 1997, 2003). It has good internal consistency, and each subscale includes items assessing minimization and denial, with all CTQ items reported on a 5-point scale ranging from “never true” to “very often true.” The scale requires approximately a sixth-grade reading level to complete. For specific scores included in our analysis, we used the total scores for the CTQ and the 3 abuse subscales measuring the history of sexual, emotional, or physical abuse for our analyses. For our primary analyses, we generated present/absent ratings based on the criteria of Walker et al. (1999). These cutoffs include scores of 8 or more on sexual and physical abuse and 10 or more on emotional abuse. Because prior research has shown that the neglect items of the CTQ do not correlate well with the 3 abuse types studied here and with independently collected data on exposure to violence (Liebschutz et al., 2018), we focused on sexual, emotional, and physical abuse and excluded neglect data in the analyses.

C-Reactive Protein

—Blood samples were collected at the time of admission during the physical examination from the antecubital vein, and CRP levels were measured at a commercial laboratory using the high-sensitivity assay for CRP. For CRP, we used the total score and generated 3 different subclassifications of blood levels of CRP: <1.0 mg/L, 1.0–3.0 mg/L, and >3.0 mg/L. For reference, any CRP levels >3 mg/L reflect abnormally high/disease-related levels of peripheral inflammation (Pepys and Hirschfield, 2003).

Body Mass Index

—We calculated BMI from height and weight measured during the admission physical examination when the blood sample for measurement of the CRP was also collected. As with the other variables, we collected BMI as a continuous variable and divided participants into low or normal weight (BMI < 24.9), overweight (BMI = 25–29.9), and obese (BMI ≥ 30) categories.

Data Analyses

Several analyses were performed. We examined the categorical associations between present/absent ratings on the 3 CTQ abuse variables (i.e., sexual, emotional, and physical abuse) and the classes of BMI and CRP measures. As a secondary analysis, we computed intercorrelations between the full range of scores on the 3 different abuse subscales, total scores on the CTQ, and the full range of scores on BMI and CRP using nonparametric Spearman Rank correlations. Finally, in the event of significant correlations between BMI and CRP and any of the categorical variables of abuse, we planned to use linear regression analysis to determine whether the associations between CRP and BMI and abuse type were independent or if elevations in CRP were not explained by increased BMI. We also examined group differences in all 4 clinical rating scales with t tests for the 3 abuse subtypes on a present/absent basis. We also created a composite clinical severity scale by standardizing all the scales and creating an aggregate score based on the average of the z scores for symptom severity for the individual scales.

We performed a power analysis to identify the size of correlations that could be meaningfully detected. Using the power analysis module of SPSS version 28, we determined that a Pearson correlation of r = .17 could be detected with our sample size of 280 at P < .05 and power = .80. Further, in terms of regression analyses, the power analysis suggested that we could detect an overall regression partial correlation of .18 at P < .05 and power >.80. Thus, we had enough power to detect correlations that share a 3% variance between variables and would be at the lowest range of possible clinical significance.

Results

Descriptive information on the patient population demographics and biometrics data are presented in Table 1, and diagnostic information is presented in Table 2. Participants were similar in self-reported gender distribution (with an average age of 27.2 years), and the sample was 84% White. Most were single and never married, and the principal diagnoses were in the mood disorders category. In addition, 46% had a comorbid substance use disorder, 42% were of average weight or less, and 28% were obese, with the high CRPs in the >3.0 range being present in 26% of the participants.

Table 1.

Demographics and Biometrics of Participants

	Total n = 280
	Variables
	Age
Mean	27.24
SD	10.12
	BMI
Mean	27.45
SD	6.86
	Blood CRP levels
Mean	3.18
SD	4.94
	CRP categories (%)
Values <1.0	50
Blood levels 1–3.0	24
Blood levels >3	26
	Length of stay
Mean	67.59
SD	31.94
	Self-reported gender (%)
Male	47
Female	53
	Racial status (%)
White	84
Black	6
Asian	4
Mixed	6
	Marital status (%)
Single, never married	75
Married	20
Divorced or separated	5
	BMI categories (%)
Normal/under	42
Overweight	30
Obese	28

Abbreviations: BMI, body mass index; CRP, C-reactive protein.

Table 2.

Diagnostic Information

Sub-types	Primary diagnosis at ­admission (%)
Major depression	59
Bipolar disorder	25
Anxiety	5
Psychosis	11
Psychotic disorders (31 cases in total)
Schizophrenia	11
Schizoaffective	18
Schizophreniform	2
Substance use disorder % (present in n = 129, or 46% of population)
Alcohol	45
Cannabis	50
Cocaine	9
Hallucinogens	5
Opiates	4
Stimulants	4
Sedatives	7
Solvents	6

Table 3 presents the means and SDs for the severity scores for abuse categories and the clinical variables and yes/no classifications based on our classification strategy: 21% for sexual abuse, 53% were positive for emotional abuse, and 27% for physical abuse. Furthermore, 27 of 280 cases (9.6%) were positive for all 3 forms of abuse. In comparison, both sexual abuse and emotional abuse (but not physical abuse) were positive for 17 of the 280 cases (6%), and both sexual abuse and physical abuse (but not emotional abuse) were present in only 2 cases (<1.0%).

Table 3.

Childhood Trauma and Clinical Symptom Scores

	Sample size (n)	Mean (% abuse history in sample)	SD
CTQ total	280	65.26	18.53
Physical abuse	280	7.37 (27)	4.53
Emotional abuse	280	11.78 (53)	6.33
Sexual abuse	280	7.43 (21)	5.75
Montgomery-Asberg depression	187	22.78	8.63
Hamilton Anxiety Rating Scale	172	19.31	8.21
Brief Psychiatric Rating Scale	40	46.8	11.89
Young Mania Rating Scale	33	8.58	6.99

Abbreviation: CTQ, Childhood Trauma Questionnaire.

Table 4 presents the categorical analyses of the abuse subtypes yes/no classification and BMI and CRP groupings. As seen in the table, for sexual abuse, but not emotional or physical abuse, there was a significantly higher number of participants with very high-level CRP scores (>3.0 mg/L; see also Figure 1). Obesity was not associated with the categorical presence of any of the 3 types of abuse. We also examined other possible patient characteristics related to the 3 different forms of abuse. We found that gender was not associated with any of the 3 abuse subtypes (all X² (1) < 1.91, all P > .39), and total scores on the CTQ did not differ as a function of gender (t₍₂₇₈₎ = .35, P = .73). Also, length of stay did not correlate with CTQ total scores (ρ= −.07, P = .314). In addition, the occurrence of a substance use disorder diagnosis did not differ as a function of the presence of any of the 3 types of abuse (all at X² (1) < 2.10, all P > .15). Further, the presence of alcohol use disorders was not related to CRP status (low, medium, high; X² (2) = .42, P = .82).

Table 4.

Associations Between Abuse Types and BMI and CRP Classifications

Abuse history	BMI category			X2	P
	Average	Overweight	Obese
Emotional abuse
% Yes	40	30	30	0.034	.99
% No	40	30	30
Physical abuse
% Yes	47	28	26	1.13	.57
% No	38	31	31
Sexual abuse
%Yes	32	24	44	5.19	.075
% No	43	32	26
CRP classifications
Abuse history	<1.0	1.0–3.0	>3.0	X2	P
Emotional abuse
% Yes	48	27	25	2.17	.5
% No	54	20	26
Physical abuse
% Yes	51	25	24	0.24	.89
% No	50	26	24
Sexual abuse
%Yes	42	20	38	6.37	.04
% No	53	25	22

Abbreviations: BMI, body mass index; CRP, C-reactive protein.

Figure 1.

Partial correlation between Childhood Trauma Questionnaire (CTQ) sexual abuse severity scores and body mass index (BMI) with C-reactive protein (CRP) values. The stepwise regression revealed that sexual abuse was significantly correlated with CRP values after inclusion of BMI in the model.

Tables 5 and 6 presents intercorrelations between the CTQ variables, their correlations with BMI and CRP, and group differences (abuse present/absent) in the 4 clinical variables, respectively. All abuse variables were significantly intercorrelated with Spearman’s rank correlation. We found correlations between childhood abuse type and CRP to be significant only for sexual abuse severity–CRP correlations (ρ= .168, P = .005), consistent with the analyses presented above regarding present/absent status for sexual abuse and being in the most elevated group for CRP. In contrast, and consistent with the group-based analyses, we did not find any significant correlations between CRP and the severity of emotional (ρ= 0.038, P = .530) or physical (ρ= −0.036, P = .549) abuse. In addition, BMI did not correlate with a history of sexual abuse (ρ= 0.13, P = .07), emotional abuse (ρ= 0.02, P = .252), or physical abuse (ρ= 0.02). Furthermore, group differences based on presence/absence of childhood abuse subtype in relation to symptom ratings were observed only for Hamilton Anxiety scores, wherein sexual and physical abuse history, but not the history of emotional abuse, was associated with significantly higher baseline scores of anxiety. Although our sample of participants was relatively evenly divided between males and females and abuse type did not differ by gender, sexual abuse, in particular, could have gender-specific correlates. To assess if our observed association between childhood abuse type and CRP levels are differentially correlated across females and males, we conducted Spearman’s correlation analyses separately between CRP and abuse type for female and male participants. Whereas we found no significant correlations between emotional abuse and CRP in females (n = 149, ρ= −0.042, P = .609) and males (n = 128, ρ= 0.064, P = .471) or between physical abuse and levels of CRP females (ρ= −0.053, P = .519) and males (ρ= 0.011, P = .901), we did find a significant correlation between sexual abuse history and CRP levels in females (ρ= 0.25, P = .003) but not in males (ρ= 0.05, P = .602. A formal comparison of the significance of the difference is not possible because we used Spearman correlations, and there is no accepted z-score transformation for rho to perform these tests.

Table 5.

Associations of CTQ Scores, Clinical Variables, CRP, and BMI Based on Spearman Correlations

	Emotional abuse	Physical abuse	Sexual abuse	CRP	BMI
CTQ total	.59**	.83**	.77**	−0.02	0.04
Emotional abuse	—	.66**	.40**	−0.06	0.02
Physical abuse	—	—	.56**	−0.07	0.04
Sexual abuse			—	.17*	0.13
CRP				—	.52**

Note. *P < .01; **P < .001

Table 6.

Group differences in clinical variables as function of abuse status

Clinical variable	Emotional abuse		Physical abuse		Sexual abuse
	t	P	t	P	t	P
Montgomery-Asberg Depression	0.91	.36	1.46	.07	0.21	.82
Hamilton Anxiety Rating Scale	1.92	.06	2.57	.01	2.53	.01
Brief Psychiatric Rating Scale	0.93	.36	1.67	.1	1.12	.26
Young Mania Rating Scale	1.29	.2	1.08	.28	0.6	.53
Composite Score	1.41	.16	0.69	.49	1.67	.1

Abbreviations: BMI, body mass index; CRP, C-reactive protein; CTQ, Childhood Trauma Questionnaire.

Partial correlation with CRP values.

Depression, mania, and psychotic symptoms indexed by the BPRS did not significantly differ between the different abuse groups, and the composite score of BPRS did not vary as a function of abuse history. Finally, the presence of PTSD comorbidity (found in 84 of the 280 participants, with 58 comorbid for major depressive disorder (MDD) and PTSD) was significantly associated with total CTQ scores. Those with comorbid PTSD had higher scores on the total CTQ (P = .016, t₍₂₇₈₎ = 2,42) as well as emotional (P < .001, t₍₂₇₈₎ = 4.12), physical (P = .003, t₍₂₇₈₎ = 3.03), and sexual abuse (P < .001, t(278) = 5.32). However, PTSD comorbidity was not related to either CRP levels (P = .47, t₍₂₇₈₎ = .49) or BMI (P = .12, t₍₂₇₈₎ = .73).

The severity of sexual abuse was associated with CRP on both dichotomous and continuous analyses, so, despite the lack of correlation between BMI and CRP, we used a regression analysis to evaluate whether the relationships between CRP, BMI, and sexual abuse were overlapping. We predicted CRP scores in the entire sample, entering BMI in the first block of regression analysis and then entering the 3 abuse severity scores in the second block in a stepwise model. The overall regression analysis was significant at F_(2,277) = 10.80, P < .001. Specifically, the BMI scores entered first in the regression accounting for 5% of the variance in CRP scores (partial correlation: r = .22, t = 2.75, P < .001). The severity of sexual abuse accounted for an additional 3% of the variance (Partial correlation: r = .17, t = 2.09, P = .031) (see Figure 1). In contrast, the severity scores for physical and emotional abuse did not predict CRP levels after the entry of BMI in the regression model (both t < 1.51, P > .13).

Discussion

In this study, we examined the relationship between the childhood trauma subtypes and later-life peripheral inflammation as assessed with CRP levels, and we found a significant influence of childhood sexual abuse on CRP. On a present/absent basis, 53%, 27%, and 21% of the participants reported childhood experiences of emotional, physical, and sexual abuse, respectively. These results underscore the high prevalence of early-life trauma in individuals receiving treatment for psychiatric disorders. Furthermore, because this was a rehabilitation facility and all participants were previously treated without recovery, these individuals comprising the current study sample presented with more refractory disease.

We found that elevations of CRP correlated significantly with only the presence of a history of sexual abuse. This observed relationship between sexual abuse and CRP scores could not be explained by our acute mood and psychotic symptoms measures, because the severity of depression, mania, and psychotic symptoms did not differ between childhood trauma subtypes. In addition, the severity of anxiety at the time of admission was also found to be elevated in participants with a physical or sexual abuse history but not emotional abuse.

Of interest, the presence of a history of sexual abuse was not as strongly related to obesity or general mental illness diagnoses. Notably, except for childhood sexual abuse history, childhood abuse did not show a significant relationship, on a continuous or dichotomous basis, with later-life circulatory CRP levels. Our current finding aligns in part with a previous study showing a trend towards a statistically significant link between childhood sexual and emotional but not physical trauma abuse with indices of physiological health in a middle-aged female-only cohort (Matthews et al., 2014). Our findings also support another recent longitudinal study that followed children and their parents from when the children were ages 9–18 years and then sampled their inflammatory CRP levels between ages 18 and 23, finding that a history of childhood sexual abuse and bullying victimization were the only 2 abuse types associated with later-life elevated CRP levels (Lob et al., 2022). Our results also lend credence to the suggestion that the traumatic nature of childhood sexual abuse renders it a far-reaching negative risk factor for later-life poor health outcomes (Matthews et al., 2014; Kiecolt-Glaser, 2015; Baumeister et al., 2016; Scheffers et al., 2017; Ananthan and Lyon, 2020; Lippard and Nemeroff, 2020; Pitharouli et al., 2021). Together, our results suggest that more than emotional and physical abuse, childhood sexual abuse may have an enduring influence on later life inflammation. Additional gender-specific correlations revealed that significant correlations between the 2 variables supported our findings of the selective association between childhood sexual abuse and CRP only in female individuals. This may suggest that childhood sexual abuse disproportionately affects females or that female individuals might have longer-lasting inflammatory repercussions to early-life sexual trauma than males. Future studies with larger sample sizes and longitudinal sampling of more inflammatory markers (in addition to CRP) and more subjective measures are needed to address the biological mechanisms related to possible gender-related differences in the link between childhood sexual abuse and lifetime inflammatory marker levels.

Of the 280 patients whose data were examined, 84 had comorbid PTSD, 58 were diagnosed with primary MDD, and 20 had a bipolar spectrum disorder. It may seem surprising that primary diagnoses of PTSD were not more common in our sample, given the pervasive history of childhood trauma, which was present in 65.26% of our participants. As noted above, childhood abuse combined with recent traumatic experiences appears to increase the risk of the near-term development of PTSD. However, adversity and trauma exposure in childhood are also clearly linked to mood disorders (Jabbi and Nemeroff, 2019; Lippard and Nemeroff, 2020), especially MDD (Gordon et al., 2020; Teicher et al., 2022) and bipolar disorder (Gordon et al., 2020; Teicher et al., 2022), leading some to suggest that the field of psychiatry needs to include childhood trauma–induced disorder as a special diagnostic category (Gordon et al., 2020; Teicher et al., 2022), much like proposals for transdiagnostic consideration of suicidal ideation and behavior.

Although our study cannot explain the mechanisms underlying the observed association between childhood sexual abuse (relative to emotional and physical abuse) history and later-life chronic inflammation as measured with CRP levels, a few possible mechanisms warrant further study. For instance, childhood sexual abuse is well-documented to be particularly pernicious in risk for later life psychopathology (Kiecolt-Glaser, 2015; Baumeister et al., 2016; Ananthan and Lyon, 2020; Pitharouli et al., 2021). Indeed, a compounding effect of childhood sexual abuse history (which is exceptionally emotional and physical by nature) can lead to elevated risk for later-life CRP and BMI levels, especially in mood disorders (Moraes et al., 2017; Pitharouli et al., 2021). Thus, the fact that 84% of our sample have mood disorders (i.e., 59% major depression and 25% bipolar disorder), the presence of childhood sexual abuse could influence our sample’s later-life inflammation and mental health status in a plethora of ways. One possible mechanism is that the repercussions of childhood sexual abuse leading to later-life mood disorders and associated fluctuations in BMI can cause an imbalanced ratio of pro-inflammatory interleukin-6 to anti-inflammatory interleukin-10 cytokines (Kim et al., 2007). Such inflammatory imbalances are likely compensatory mechanisms designed to mitigate the 1-sided increase in pro-inflammatory markers in sickness states such as the documented childhood sexual trauma–related later-life episodic mood disorder symptoms and comorbid conditions in our study sample (Tsigos and George, 2002; Kim et al., 2007). Future studies designed to assess the relationship between specific types of early-life trauma and later-life disorder/symptom-associated inflammatory markers will inform mechanisms.

It is important to note that, although our findings are based on a sample of well-characterized psychiatric patients in a relatively controlled environment, the following limitations deserve mention. Our sample size of 280 participants is not large, and minorities are under-represented, as most of the samples are Caucasian. Additionally, this rehabilitation community consists of only private insurance holders or individuals with the ability to pay out of pocket, which reduces the socio-economic heterogeneity of the sample, making our study sample not representative of the US population in terms of normal distribution of socio-economics and racial makeup. However, these findings do suggest that even children from higher SES strata are still commonly vulnerable to various abuse experiences in childhood.

The average sample age of 27.2 years is also relatively young. Seventy-five percent of the participants were either single or never married at the time of the study, which is not surprising given the relatively young age range of the sample. Psychiatric diagnoses were unevenly distributed, with major depression predominating and psychosis relatively rare. All participants were receiving treatment for persistent psychiatric morbidities, which might be associated with a lack of difference in the prevalence of the abuse types. The CRP measures could be impacted by medication use; because our current CRP data were collected at the time of admission, we could not fully control for prior medication. A lack of reliable data on medication use before admission means the possible effects of medication use before admission to the facility could not be accounted for. However, the specific correlation between CRP and only 1 of the 3 abuse subtypes argues against generic effects of medication treatment or other systemic illness because it is unclear how this bias could be related to a single type of early childhood abuse. Forty-six percent of our sample had a history of comorbid substance use disorder even though none of these variables showed any significant relationship with CRP or BMI status. Finally, although our focus on CRP measures from peripheral blood has yielded important findings, the interpretations of our results showing selective inflammatory CRP correlates for sexual abuse are limited by the lack of interactive biomarkers such as brain imaging measures (Sheffield et al., 2013), glucocorticoid metrics of the hypothalamic pituitary adrenal (HPA)-axis responses (Stein et al., 1997; Sapolsky et al., 2000), and transcription factor activity such as nuclear factor-κB pathway responses (Pace et al., 2012). Considering our findings of interaction between sexual abuse history and CRP levels, additional measures of glucocorticoid levels such as cortisol at admission could reveal further mechanistic insights given the role of glucocorticoids as transcription factors that mediate immune/inflammatory responses and in influencing both carbohydrate and blood glucose metabolic functions during stressful and traumatic life events (Sapolsky et al., 2000).

Future studies need to recruit more representative samples across a wider age range spanning socio-economic and racial groups, including younger individuals, where the abuse may have been even more proximal, and older participants to examine the lifelong persistence of these effects. Finally, the relationship between treatment response and measures of inflammation is an important issue that needs to be addressed with follow-up data, which are now being collected at this research site.

Conclusion

The high prevalence rate of childhood trauma in this psychiatric population and the link between abuse history and later-life negative consequences was suggested by our findings of a specific association between childhood sexual abuse and early-adulthood inflammatory status. This finding is a compelling example of how childhood environmental adversity can have lingering negative physiological consequences at a critical developmental period. Future studies need to include more representative samples and examine how early-life sexual trauma can negatively affect physical health outcomes.

Conflict of Interest

M.J.: None. P.D.H.: Dr Harvey has received consulting fees or travel reimbursements from Alkermes, ANeuroTech (division of Anima BV), Bio Excel, Boehringer Ingelheim, Karuna Pharma, Minerva Pharma, and Sunovion Pharma during the past year. In addition, he receives royalties from the Brief Assessment of Cognition in Schizophrenia (Owned by Verasci, Inc). He is Chief Scientific Officer of i-Function, Inc. R.J.K.: Dr Kotwicki is a full-time employee of Skyland Trail. C.B.N. Consulting (last 12 months): ANeuroTech (division of Anima BV), Signant Health, Sunovion Pharmaceuticals, Inc., Janssen Research & Development LLC, Magstim, Inc., Navitor Pharmaceuticals, Inc., Intra-Cellular Therapies, Inc., EMA Wellness, Acadia Pharmaceuticals, Axsome, Sage, BioXcel Therapeutics, Silo Pharma, XW Pharma, Neuritek, Engrail Therapeutics, Corcept Therapeutics Pharmaceuticals Company. Stockholder: Xhale, Seattle Genetics, Antares, BI Gen Holdings, Inc., Corcept Therapeutics Pharmaceuticals Company, EMA Wellness, TRUUST Neuroimaging. Scientific Advisory Boards: ANeuroTech (division of Anima BV), Brain and Behavior Research Foundation (BBRF), Anxiety and Depression Association of America (ADAA), Skyland Trail, Signant Health, Laureate Institute for Brain Research (LIBR), Inc., Magnolia CNS, Heading Health, TRUUST Neuroimaging. Board of Directors: Gratitude America, ADAA, Xhale Smart, Inc.

Patents

-Method and devices for transdermal delivery of lithium (US 6,375,990B1)

-Method of assessing antidepressant drug therapy via transport inhibition of monoamine neurotransmitters by ex vivo assay (US 7,148,027B2)

Funding

MJ received funding from the National Institutes of Health (NIH), #R21MH115326. PDH received funding from NIH, R01AG051346. CBN received funding from NIH, #R01MH117293.