Select Your Interests

Elevated Plasma Inflammatory Markers in Individuals With Intermittent Explosive Disorder and Correlation With Aggression in Humans

A substantial body of research demonstrates bidirectional relationships between the brain and behavior with immune function^1,2 and vice versa.³ For example, depressed patients consistently display evidence of elevated inflammatory cytokine levels^4,5 and the therapeutic administration of inflammatory cytokines is associated with increased depressive symptoms.^6,7

In addition, animal^8-11 and human^12-19 studies suggest that behavioral traits related to hostile, angry, and aggressive disposition are associated with elevations in inflammatory markers. Defensive rage in cats is associated with elevated levels of interleukin 1β (IL-1β) and IL-6 and blocking IL-1β activity reduces these effects^8-10 and mice deficient in inflammatory cytokine receptors fail to exhibit aggressive and defensive behavior even when threatened.¹¹ Studies in humans, also, suggest a direct relationship between C-reactive protein (CRP) and hostile,^12,13 angry,¹⁴ and aggressive dispositions^15,16; elevations of IL-6 are associated with the same variables.^16-18,20

While work in healthy humans suggests a direct relationship between hostility, anger, and aggressive disposition and inflammatory markers, to our knowledge, no study has been performed in psychiatric patients with prominent histories of impulsive aggressive behavior. In this study, we hypothesized that plasma levels of inflammatory markers will be elevated in individuals with a current diagnosis of intermittent explosive disorder (IED), a disorder of impulsive aggressive behavior,^21,22 and that these markers will correlate directly with measures of aggression (and/or impulsivity) in healthy participants and psychiatric patients. We assessed levels of plasma CRP, because it is an acute-phase reactant released in the presence of inflammation,²³ and plasma IL-6, because it is an inflammatory cytokine²⁴; both markers have been shown to be elevated as a function of hostility.^16-18,20

One-hundred ninety-seven physically healthy participants were recruited from clinical settings and through newspaper advertisements, seeking out individuals who reported psychosocial difficulty related to 1 or more Axis I or II conditions or who had little evidence of psychopathology. All participants gave written informed consent as approved by the University of Chicago institutional review board. Participants with bipolar disorder, schizophrenia, or mental retardation were excluded. Medical health was documented by comprehensive medical history and examination, which included a screen for drugs of abuse (all participants tested negative).

Axis I and Axis II diagnoses were made byDSM-IV criteria²⁵; IED diagnoses were made by research criteria.²² Research assessments were performed by individuals with master’s/doctoral degrees in clinical psychology with interrater (κ) reliability ranging from 0.79 to 93 (mean [SD], 84 [0.05]) across mood, anxiety, substance use, impulse control, and personality disorders. Final diagnoses were assigned by previously described best-estimate consensus procedures,^26-28 using information from (1) Structured Clinical Interview forDSM-IV Axis I Disorders²⁹; (2) Structured Interview forDSM-IV Personality: SIDP-IV³⁰; (3) clinical interview by a research psychiatrist; and (4) review of all available clinical data.

Sixty-nine participants met criteria for current IED, 61 for other current/lifetime Axis I and/or Axis II disorder (psychiatric controls [PC]), and 67 without evidence of anyDSM-IV disorder (healthy controls [HC]). Among participants with IED and PC, 112 had current history of an Axis I disorder, 125 had lifetime history of an Axis I disorder, and 105 had an Axis II personality disorder. About one-third (35%) of participants with IED and PC had history of psychiatric treatment. Specific diagnoses for participants with IED and PC are listed in the eAppendix in theSupplement.

Aggression was assessed with the Aggression Scale from Life History of Aggression³¹ (LHA) and the Buss-Perry Aggression Questionnaire³² (BPAQ). Impulsivity was assessed with the Life History of Impulsive Behavior³³ (LHIB) and Barratt Impulsiveness Scale³⁴ (BIS-11). The LHA and LHIB assess the number of times a person has engaged in aggressive, or impulsive, behavior in his or her life and the BPAQ and BIS-11 assess a person’s disposition to act aggressively, or impulsively, as a personality trait. Life history of suicidal behavior was assessed during the Structured Clinical Interview. Other assessments included the Beck Depression Inventory³⁵ for state depression, Life Experiences Survey³⁶ for stressful life events over the past 6 months, and Eysenck Personality Questionnaire³⁷ for control dimensions of general personality. The Global Assessment of Function²⁵ Scale served as the variable for psychosocial functioning. Racial data, collected by diagnostic assessors, reflected self-identified racial characteristics of participants.

Participants were free of all medications for at least 4 weeks. Whole blood, anticoagulated with EDTA, was obtained between 9 and 11am through venipuncture of a forearm vein. Plasma was processed after centrifugation and stored at −80°C until assay. The CRP and IL-6 levels were detected by commercially available (R&D Systems) enzyme-linked immunosorbent assay kits and assayed according to the manufacturer’s instructions. Sensitivity and coefficient of variation were 0.035 mg/L and less than 6% for CRP and 0.70 pg/mL and less than 6% for IL-6, respectively. Samples were run together in the same assay and levels reported represent the mean of the duplicates.

Comparisons between groups were performed byt test, analyses of variance and covariance, and χ² tests. Correlational analyses were conducted by parametric and nonparametric methods as necessary. An α of .05 denoted statistical significance. Plasma CRP levels were normally distributed; IL-6 levels were not and were log-transformed for analysis. Given the known relationship between body mass index, age, state depression (Beck Depression Inventory), psychological stress (Life Experiences Survey), and inflammation,^3-5 data were first analyzed without these variables and then with these variables as covariates. Selected demographic and lifestyle variables were also used as covariates. Finally, composite variables for “aggression” and “impulsivity” were created in a data-reduction step by taking the average of each participant’sz scores for the primary behavioral measures (LHA and BPAQ; LHIB and BIS-11).

Participants did not differ in distribution of sex or race (Table 1). The groups did differ in age and Hollingshead socioeconomic status score. Differences in age were due to a younger age among HC vs participants with IED and PC; differences in socioeconomic status were due to a greater proportion of participants with higher socioeconomic status (eg, categories I and II) among HC (76%) and PC (77%) vs participants with IED (51%). Neither mean body mass index nor rate of obesity (ie, body mass index >30 [calculated as weight in kilograms divided by height in meters squared]) differed across groups. The rate, or degree, of current cigarette smoking also did not vary across groups. While the groups did not differ in rate of current social drinking, participants with IED reported a higher mean number of drinks per day and a higher rate of consuming more than 2 drinks per day (HC = PC<participants with IED). As expected, the groups differed in LHA/BPAQ aggression, LHIB/BIS-11 impulsivity, Beck Depression Inventory, and Life Experiences Survey scores (participants with IED>PC>HC) and in Global Assessment of Function scores (participants with IED<PC<HC).

Multivariate analysis of variance revealed a significant difference among the groups with both inflammatory markers (Wilks λ = 0.43;F_6,374 = 32.66;P < .001; CRP:F_2,189 = 21.67;P < .001; log IL-6:F_2,189 = 70.85;P < .001). In both cases, participants with IED displayed higher inflammatory marker levels than either HC or PC (Figure 1). Addition of the a priori covariates of body mass index, age, state depression, and recent psychological stress did not change this result (Table 2).

Analysis of variance of plasma CRP and log IL-6 levels revealed significant IED vs PC differences for participants with IED in the context of other current Axis I disorders (CRP:F_1,124 = 8.54;P = .004; log IL-6:F_1,122= 190.09;P < .001), participants with IED in the context of lifetime Axis I disorders (CRP:F_1,122 = 6.25;P = .01; log IL-6:F_1,120 = 147.07;P < .001), and participants with IED in the context of Axis II personality disorder cluster (CRP:F_1,126 = 6.90;P = .01; log IL-6:F_1,124 = 157.69;P < .001). Analysis of variance of participants with IED and PC with current Axis I disorders (ie, removing participants with only life history of Axis I disorder) did not change the group results (mean [SD], CRP: participants with IED = 2.36 [0.70] vs PC = 1.48 [1.35] mg/L;F_1,109 = 20.36;P < .001; log IL-6: participants with IED = 0.23 [0.29] vs PC = −0.90 [0.55] mg/L;F_1,109 = 225.97;P < .001). Adding history of psychiatric treatment to the analysis did not change these group results.

Neither levels of CRP nor log IL-6 differed as a function of sex (mean [SD], CRP: male = 1.65 [1.07] vs female = 1.83 [1.18] mg/L;t₁₉₅ = 1.11;P = .27; log IL-6: male = −0.21 [0.66] vs female = −0.31 [0.076] pg/mL;t₁₉₃ = 0.95;P = .34). Among all demographic variables, overall, only age covaried at a trend level of significance with CRP (multiple regression:F_4,192 = 2.73;P = .06; partialr = 0.18;P < .02), and only race covaried significantly with log IL-6 levels (multiple regression:F_4,190 = 5.79;P < .001; partialr = 0.30;P < .001). The influence of age on the inflammatory markers was already accounted for as an a priori covariate in the earlier analyses (Table 2); adding race as a covariate did not affect the group results for log IL-6 levels (F_2,186 = 12.93;P < .001). While the groups did not differ in the proportion of obese participants, the presence of obesity was associated with elevated levels of CRP (mean [SD], 2.33 [1.15] vs 1.59 [1.08] mg/L;t₁₉₅ = 3.85;P < .001) and log IL-6 (mean [SD], 0.01 [0.64] vs −0.34 [0.72] pg/mL;t₁₉₃ = 2.87;P = .005). Removal of obese participants did not change the group results (CRP:F_2,153 = 24.46;P < .001; log IL-6:F_2,151 = 57.81;P < .001). While the groups also differed in mean drinks per day (and in the rate of participants consuming >2 drinks per day), this variable had no effect on levels of CRP (F_1,193 = 0.83;P = .37) or log IL-6 (F_1,191 = 0.69;P = .41) when added to the statistical model; removal of participants reporting more than 2 drinks per day did not change the group results (CRP:F_2,181 = 21.14;P < .001; log IL-6:F_2,179 = 61.74;P < .001). The smoking variable, current packs per day, also had no effect on levels of CRP (F_1,192 = 0.80;P = .37) or log IL-6 (F_1,190 = 0.40;P = .85) when added to the statistical model; removal of current smokers did not change the group results (CRP:F_2,151 = 18.95;P < .001; log IL-6:F_2,150 = 55.29;P < .001).

Across all participants, the core features of IED, aggression and impulsivity, displayed a direct relationship with levels of both CRP and log IL-6: composite aggression: CRP:r = 0.39; n = 176;P < .001; log IL-6:r = 0.37; n = 176;P < .001; composite impulsivity: CRP:r = 0.31; n = 177;P < .001; log IL-6:r = 0.26; n = 175;P = .001. Plasma CRP and log IL-6 were also significantly correlated in these participants (r = 0.47; n = 175;P < .001). While composite aggression and composite impulsivity scores were highly correlated in these participants (r = 0.66; n = 166;P < .001), multiple regression analysis (F_2,163 = 16.25;P < .001;R = 0.41;R² = 0.16) revealed a unique contribution for composite aggression (partialr = 0.27;t₁₆₃ = 3.57;P < .001), but not for composite impulsivity (partialr = 0.08;t₁₆₃ = 0.97;P = .33) for CRP levels. The same was true for log IL-6 levels (F_2,161 = 13.45;P < .001;R = 0.38;R² = 0.13; composite aggression: partialr = 0.31;t₁₆₁ = 4.14;P < .001; composite impulsivity: partialr = −0.03;t₁₆₁ = 0.41;P = .69). Addition of the a priori covariates (body mass index, age, Beck Depression Inventory score, and Life Experiences Survey score), individually and together, continued to reveal significance for correlations between composite aggression and CRP and log IL-6 (Table 3). Examination of the demographic and lifestyle variables did not change the results.

Within the composite aggression variable, multiple regression (F_2,175 = 19.54;P < .001;R = 0.43;R² = 0.17) revealed a unique contribution of LHA aggression (partialr = 0.32;t₁₇₅ = 4.52;P < .001), but not of BPAQ aggression (partialr = 0.00;t₁₇₅ = 0.06;P = .95), for plasma CRP. The same was true for log IL-6 (F_2,173 = 14.49;P < .001;R = 0.38;R² = 0.13; LHA: partialr = 0.22;t₁₇₃ = 3.02;P = .003; BPAQ: partialr = 0.09;t₁₇₃ = 1.18;P = .24). Moreover, addition of LHA aggression scores to the analysis of covariance model (F_6,354 = 19.78;P < .001) reduced differences between participants with IED vs PC vs HC to a nonstatistically significant trend for CRP (F_2,179 = 2.40;P = .09) and reduced the IED vs HC difference by 49%. While group differences remained significant for log IL-6 (F_2,179 = 45.25;P < .001), the IED vs HC difference for log IL-6 was reduced by 47%. The scatterplots for the correlation between LHA aggression and CRP and log IL-6 levels are displayed inFigure 2.

Aggression by LHA correlated significantly with both inflammatory makers alone and when examined together with Eysenck personality variables in multiple regression analyses (CRP: partialr = 0.34;t₁₇₃ = 4.79;P < .001; log IL-6: partialr = 0.30;t₁₇₁ = 4.06;P < .001). Coefficients for partial correlations between Eysenck personality variables, with each plasma inflammatory marker, were all nonsignificant.

Among participants with IED and PC, life history of suicide attempt was associated with a trend for higher CRP levels (F_1,128 = 2.99;P = .09) and a significant difference for higher log IL-6 levels (F_1,126 = 8.26;P = .005). Participants with life history of suicide attempt also had higher LHA aggression scores, (mean [SD], suicide attempt: 18.6 [7.3] vs no suicide attempt: 13.9 [7.1];t₁₂₀ = 2.51;P = .01) and adding LHA aggression scores to the analysis of covariance model eliminated these differences for both CRP (F_1,118 = 0.02;P = .89) and log IL-6 (F_1,116 = 0.06;P = .80).

To our knowledge, this is the first study to examine plasma inflammatory markers in a psychiatric sample that includes well-characterized participants with current history of recurrent, problematic, impulsive aggression, diagnosed with IED, and to examine behavioral, nonpersonality-based measures of aggression and impulsivity. In this sample, plasma markers of the inflammatory process displayed a direct relationship with aggression and were elevated, specifically, in participants with IED. Accordingly, this suggests that objective signs of systemic inflammation^23,24 are directly associated with aggression and are present in individuals with IED compared with both HC and PC. While one action of IL-6 is to stimulate production of CRP from the liver,²³ and the 2 markers were correlated in this study, the observation that both independently correlated with aggression suggests that both have unique relationships with aggression. Collateral support for this possibility comes from a population-based study³⁸ reporting significant elevations for participants with IED in the risk of conditions with inflammatory components (eg, coronary artery disease, cerebrovascular disease, arthritis, and peptic ulcer). Finally, the relationship with history of suicidal behavior, in this particular sample, may be fully mediated by the relationship between aggression and inflammatory markers.

These results are not likely due to other relevant factors. Participants were physically healthy, free of systemic illness or physical injury, free of any medication for at least 4 weeks, and not abusing drugs of any kind. In addition, these results cannot be accounted for by sex because neither marker differed by sex or by differences in rates of current Axis I conditions. Removal of participants with IED and PC with only a lifetime Axis I condition did not alter these results. More importantly, these findings were not meaningfully changed when adjusted for differences in body mass index, state depression, age, or history of stressful life events This is noteworthy because each of these factors has been associated with elevation of a number of inflammatory markers.^3-5,39 These results were also not due to any influence of demographic or lifestyle factors on plasma CRP or log IL-6 levels. Results were unchanged when demographic variables were added to the statistical models; lifestyle variables were similar across groups and had no relationship with either plasma CRP or log IL-6 levels. These results were also not due to the presence of other Axis I or II conditions, other than IED, which is the categorical expression of the construct of impulsive aggression. Finally, measures of general personality dimensions did not correlate with these inflammatory markers, indicating that the relationship is with impulsive aggression, more specifically, than with general personality dysfunction.

It is possible that the elevation of inflammatory markers in participants with IED is the result of the “stress” of frequent aggressive interactions. In the present study, participants were studied at rest and no participant reported any meaningful stressor on the day of the sample collection or reported any recent physical injury. In addition, while our measure of stressful life events may not have captured potential stresses close in time to collection of the blood sample, this measure did show clear elevation of stressful life events in the participants with IED vs PC or HC, and controlling for this variable did not change the statistical significance of the findings.

These data are consistent with previous human studies. Self-assessed anger and aggressive disposition are increased in patients treated with cytokine immunotherapy.^6,7 Further, plasma CRP and IL-6 levels correlate directly with related measures of hostility,^{12-14,17,18,20} anger,¹⁴ and aggressive dispositon^15,16 in healthy adults. Greater levels of anger and/or hostility are also associated with greater production of inflammatory cytokines from blood monocytes after stimulation by bacterial lipopolysaccharide.⁴⁰ Finally, angry marital interactions have also been associated with increases in inflammatory cytokines.¹⁹

The inflammatory markers in this study appear to be associated specifically with history of actual aggressive behavior (LHA) than with the presence of aggressive tendency (BPAQ) or impulsivity as a behavior (LHIB) or as a personality trait (BIS-11). Despite the univariate correlation between BPAQ and LHIB/BIS measures and inflammatory markers, the observed relationship with both is through its shared variance with history of actual aggressive behavior (LHA). This observation does not contradict reports of similar relationships between inflammatory markers and self-assessed measures of anger and/or hostility.^{12-14,17,18,20} However, when history of actual aggressive behavior is included, other related variables may add little beyond what is associated with history of actual aggressive behavior. This observation is consistent with the findings of a recent study in which aggressive disposition (ie, BPAQ), but not cognitive (ie, “hostility”) or affective (ie, “anger”) components of antagonistic characteristics, were associated with elevations of plasma CRP and IL-6.¹⁶

While inflammatory cytokines exist in the brain, and act as neuromodulators, the markers in this study are largely produced in the periphery. However, circulating cytokines can access the brain through a number of pathways including passage through leaky regions in the blood-brain barrier, active transport through saturable transporters, activation of cells lining the cerebral vasculature that produce cytokines, and binding to receptors on peripheral afferent nerve fibers that can relay cytokine signals to relevant brain regions such as the hypothalamus and other brain structures.^41,42 Further, preclinical study has documented that inflammatory cytokines (eg, IL-1β and IL-2) administered into the medial hypothalamus or periaqueductal gray^10,43-46 increase defensive-rage aggression in cats. Accordingly, it is possible that elevations of peripheral inflammatory proteins can affect aggressive behavior through direct effects on brain regions important in the modulation of these behaviors.

The strengths of this study include a well-characterized sample, multiple validated measures of aggression and impulsivity, and a standardized approach to minimize the effect of extraneous factors on inflammatory marker levels. Limitations include the cross-sectional nature of the study and that correlational analysis cannot establish causality. While many participants had no prior psychiatric treatment, history of psychiatric treatment played no role in these findings. Finally, while our participants were physically healthy by history and examination, other relevant biomedical variables (eg, hematologic, lipid, coagulation, and complement measures) that might aid in fully ruling out existing medical conditions were not collected. Despite this, no participant had a plasma CRP level more than 4.96 mg/L. Given that the upper normal limit of plasma CRP is 10 mg/L, it is unlikely these findings could be accounted for by a known medical cause.

In summary, we report a direct relationship between 2 plasma inflammatory markers and aggression in human participants, particularly in those with IED. This relationship was not accounted for by possible demographic/lifestyle confounds including the presence of other psychiatric disorder or general personality factors other than aggression. Given that IED, a disorder of impulsive aggression, displays a 2% to 3% one-year prevalence rate in the United States,⁴⁷ and that currently available treatments bring less than 50% of those cases into remission,⁴⁸ additional strategies for the examination and intervention of human impulsive aggression are needed.

Corresponding Author: Emil F. Coccaro, MD, Clinical Neuroscience Research Unit, Department of Psychiatry and Behavioral Neuroscience, University of Chicago, 5841 South Maryland Ave, Chicago, IL 60637 ([email protected]).

Submitted for Publication: March 13, 2013; final revision received June 10, 2013; accepted July 16, 2013.

Published Online: December 18, 2013. doi:10.1001/jamapsychiatry.2013.3297.

Author Contributions:Drs Coccaro and Coussons-Read had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design:Coccaro, Coussons-Read.

Acquisition of data:All authors.

Analysis and interpretation of data:All authors.

Drafting of the manuscript:All authors.

Critical revision of the manuscript for important intellectual content:All authors.

Statistical analysis:All authors.

Administrative, technical, or material support:Coccaro, Coussons-Read.

Conflict of Interest Disclosures: Dr Coccaro reports being on the scientific advisory board of Azevan Pharmaceuticals Inc. Dr Lee reports being the recipient of a research grant from Azevan Pharmaceuticals Inc. No other disclosures were reported.

Funding/Support: This work was supported in part by National Institute of Mental Health grants RO1 MH60836, RO1 MH63262, RO1 MH66984, and RO1 MH80108 (Dr Coccaro) and a Project Pilot Grant from the University of Colorado Denver (Dr Coussons-Read).

Role of the Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.