Severe Affective and Behavioral Dysregulation in Youths Is Associated with a Proinflammatory State

1MH and LP contributed equally to the paper

Introduction

Severe affective and behavioral dysregulation, including irritability, aggression, “affective storms,” and mood instability present a challenging phenotype in a sizable number of adolescents (Schmitt, Gold, & Rauch, 2012). Since this dysregulation complex is not explicitly captured by DSM-IV-TR and ICD-10, the fifth version of the DSM now includes the diagnoses of a “disruptive mood dysregulation disorder (DMDD),” comprising children exhibiting persistent irritability and severe behavioral outbursts (APA, 2013).

To date, most researchers used the Child Behavior Checklist (Achenbach, 1991), which is one of the best-studied, empirically derived parent checklists to measure general child and adolescent psychopathology, in order to identify the mixed phenotype of severe dysregulation. This Child Behavior Checklist-Dysregulation Profile (CBCL-DP) is characterized by extreme values on the syndrome scales Anxious/Depressed (A/D), Attention Problems (AP), and Aggressive Behavior (AGG). The CBCL-DP is highly heritable (Althoff, Rettew, Ayer, Sulman, & Hudziak, 2006; Boomsma et al., 2006; Hudziak, Althoff, Derks, Faraone, & Boomsma, 2005), stable across ages (Boomsma et al., 2006) and well corroborated by replication studies (Mick, Biederman, Pandina, & Faraone, 2003). In addition, it has been found to be associated with suicidal behavior, difficulties to go to sleep and sleep through as well as a reduced need for sleep (Holtmann, Bölte, Goth, & Poustka, 2008; Holtmann, Buchmann et al., 2011; Ayer et al., 2009; Legenbauer, Heiler, Holtmann, Fricke-Oerkermann, & Lehmkuhl, 2012). About 1–2% in epidemiological samples, 6–7% in child psychiatric clinical samples, and 13–20% of children with attention-deficit/hyperactivity disorder meet the criteria for the CBCL-DP (Holtmann et al., 2007, 2008; Holtmann, Becker, Banaschewski, Rothenberger, & Roessner, 2011; Volk & Todd, 2007).

Studies on its nosological validity revealed evidence of genetic (Althoff et al., 2006; Boomsma et al., 2006; Hudziak et al., 2005), neurometabolic (Zepf, Wöckel, Poustka, & Holtmann, 2008), and endocrinologic (Holtmann, Duketis, Goth, Poustka, & Bölte, 2010) fundamentals.

The importance of immune activation processes for psychiatric disorders is increasingly recognized (Goldstein, Kemp, Soczynska, & McIntyre, 2009; A. H. Miller, Maletic, & Raison, 2009). Increases in inflammatory acute phase proteins (APPs) have been reported for a variety of mental disorders, even in the absence of an acute infection. Meta-analytic evidence suggests that major depressive disorder is associated with a proinflammatory state, including high positive APP levels, among them C-reactive protein (CRP) (Howren, Lamkin, & Suls, 2009), and decreased negative APP levels, specifically albumin (Zorilla et al., 2001). The occurrence of multiple depressive episodes in adolescents and young adults exerts the greatest effect on later CRP levels (Copeland, Shanahan, Worthman, Angold, & Costello, 2012). Studies also found evidence of immune activation in bipolar disorder (Goldstein et al., 2009), aggression (Coccaro, 2006), sleep disorders (Irwin et al., 2008), maladaptive emotion regulation (Appleton, Buka, Loucks, Gilman, & Kubzansky, 2013) and suicidality (Lindqvist et al., 2009). While most studies were conducted in adults, immune activation was also reported in youths with posttraumatic stress disorder (Pervanidou, 2008), adolescent major depression (Gabbay et al., 2009), and adolescents with a high frequency of interpersonal stressors (Fuligni et al., 2009) and childhood adversity (G. E. Miller & Cole, 2012). These diagnoses and symptoms have commonly been linked to the CBCL-DP.

While there is increasing evidence that the CBCL-DP predicts serious psychopathology in young adulthood (Biederman et al., 2009; Meyer et al., 2009; Holtmann, Buchmann et al., 2011), very little is known about the impact of the CBCL-DP on the physical health of affected youths. Elevated levels of CRP are associated with a greater risk for the later development of cardiovascular disease and other medical morbidities (Pearson et al., 2003).

Given the interest in identifying at-risk children for later health problems, the focus of the present study was to examine the link between severe dysregulation in youths and indicators of chronic inflammation. Therefore, two subgroups of a child and adolescent psychiatric sample exhibiting severe and low levels of dysregulation, respectively were compared regarding serum levels of CRP and albumin, two robustly responding biomarkers of low-grade inflammation. We hypothesized that high levels of dysregulation would be associated with higher levels of CRP and lower levels of albumin.

Method

The study comprised a sample of 133 children and adolescents (57 girls and 76 boys), aged 5.0 to 17.8 years (mean 12.1 ± 3.6 years), who had been referred to child psychiatric care at the Department of Child and Adolescent Psychiatry of the Goethe University Frankfurt/Main, Germany. The department provides child psychiatric care for a population of 660,000 inhabitants, covering a predominantly urban area. The hospital is the sole provider of inpatient care for the study area. At admission, no participant was receiving any psychotropic drugs that could potentially affect inflammatory processes.

Discharge ICD-10 diagnoses were established by experienced child and adolescent psychiatrists in weekly consensus conferences based on semistructured parent and child interviews administered by experienced clinicians (for details, see Holtmann et al., 2010). Psychosocial functioning was assessed using the German adaptation of the Children’s Global Assessment Scale (Remschmidt, Schmidt, & Poustka, 2006). The socioeconomic status was measured according to the parents’ educational and occupational achievement (Englert & Poustka, 1995).

The parents or caregivers of all participants completed the German version of the Child Behavior Checklist. In 32 subjects (24.1%), no information was available on who had completed the CBCL; 60.9% of the checklists were completed by mothers and 15.0% were filled out by fathers. Following the method of Hudziak et al. (2005), we defined the CBCL-DP score as the composite T-score of the three CBCL subscales AP, AGG, and A/D. Children with a CBCL-DP score in the definite clinical range (exceeding 225, i.e., 2.5 SDs above average) constituted the subgroup with pronounced dysregulation (CBCL-DP; n = 53). Children with CBCL-DP scores below 180 (1 SD above average or less) were regarded as psychiatric controls without dysregulation (n = 61).

No group differences emerged for gender (χ² = .095, p = .757), age (Z = –1.40, p = .164), socioeconomic status (χ² = 1.97, p = .577), global functioning (Z = –0.21, p = .830), and weight (Z = –0.77, p = .442) (Table 1 ).

Table 1. Age, weight, sex distribution, and diagnoses in patients with severe dysregulation (CBCL-DP) and psychiatric controls (CON)

	CBCL-DP (n = 51)	CON (n = 82)		p
Age (years; mean ± SD)	11.6 ± 3.5	12.4 ± 3.7	Z = –1.40	.164
Weight (kg; mean ± SD)	46.6 ± 20.5	51.1 ± 22.2	Z = –0.77	.442
Sex (male: female)	30: 21	46: 36	χ²= .095	.757
ICD-10 Diagnosis (n)
F32/F33 Depressive disorders	1	6
F40/F41 Phobia/Anxiety disorders	–	6
F42 Obsessive compulsive disorder	3	3
F43 Adjustment disorders	2	8
F50 Eating disorders	1	11
F60 Personality disorders	1	2
F84 Pervasive developmental disorders	2	3
F90.0 Hyperkinetic disorder	4	13
F90.1 Hyperkinetic conduct disorder	14	5
F91 Conduct disorder	11	8
F92 Disorders of conduct and emotions	9	2
F93 Anxiety disorders of childhood	1	4
F94 Attachment disorders	1	1
F98 Enuresis/Encopresis	1	3
Other	–	7

Table 1. Age, weight, sex distribution, and diagnoses in patients with severe dysregulation (CBCL-DP) and psychiatric controls (CON)

View as image HTML

Blood samples were taken between 8 and 10 a.m. CRP and albumin assays were performed using a latex immunoturbidimetric test (CRPLX, Roche/Hitachi; sensitivity: 0.0425 mg/dl) and a BCG procedure (ALB plus, Roche/Hitachi; sensitivity: 0.2 g/dl). Following the criterion suggested by the American Heart Association (Pearson et al., 2003), participants with high CRP values (1.0 mg/dl) suggesting an acute inflammatory response (e.g., an infection) were excluded. This left n = 117 patients with CRP and n = 85 with albumin data.

Data Analysis

Since the levels of serum CRP and albumin were not normally distributed, differences between groups were ascertained using Mann-Whitney U-tests. Group differences in the number of subjects with elevated CRP were determined using a cutoff point of 0.30 mg/dl because levels above this point place adult subjects in a high-risk group for cardiovascular events (Pearson et al., 2003). Spearman’s rank correlations were computed between CRP, albumin, and the unweighted sum of the three CBCL-DP scales (CBCL-DP score).

As suicidality and decreased sleep duration are associated both with severe dysregulation (Ayer et al., 2009; Holtmann, Buchmann et al., 2011) and proinflammatory states (Irwin et al., 2008; Lindqvist et al., 2009), APP levels were also compared between participants scoring or not scoring positive for the following items: (1) one of the two suicidal items on the CBCL (“Deliberately harms self or attempts suicide,” and “Talks about killing self”); (2) the CBCL item “sleeps less than most kids.” These items are not part of the CBCL-DP scales.

Results

In general, the mean and standard deviation values of CRP and albumin in the total sample mirror those of physically healthy European youths (Fuligni et al., 2009) (Table 2 ). Serum CRP was higher in the CBCL-DP group (0.20 ± 0.15 mg/dl) than in controls (0.15 ± 0.16 mg/dl; Z = 2.48, p = .013). Albumin was lower in the CBCL-DP group (4.69 ± 0.29 g/dl) than in controls (4.90 ± 0.31 g/dl; Z = –3.81, p < .001).

Table 2. Serum levels of CRP and Albumin in patients with severe dysregulation (CBCL-DP) and psychiatric controls (CON)

	Total sample	CBCL-DP	CON	Mann-Whitney-Test	p
Note.1n = 117; 2n = 85.
CRP (mg/dl)1	0.17 ± 0.16 range: 0.0–0.63	0.20 ± 0.15	0.15 ± 0.16	Z = 2.48	.013
Albumin (g/dl)2	4.80 ± 0.31 range: 3.90–5.80	4.69 ± 0.29	4.90 ± 0.31	Z = –3.81	<.001

Table 2. Serum levels of CRP and Albumin in patients with severe dysregulation (CBCL-DP) and psychiatric controls (CON)

View as image HTML

CRP-values in the high-risk category (≥ 0.30 mg/dl) were reached by 44.4% of CBCL-DP participants vs. 18.1% of controls (χ² = 9.52, p = .002).

CRP was positively correlated with the CBCL-DP score (r = .23, p = .013). A negative correlation was observed between albumin and the CBCL-DP score (r = –.31, p = .004).

Reduced need for sleep was more prevalent among CBCL-DP than among controls (χ² = 9.08, p = .003); participants reporting reduced need for sleep had lower albumin levels (Z = –2.58, p = .010), while CRP did not differ from controls (Z = –.75, p = .456). Suicidality was unrelated to albumin and CRP levels.

CRP and albumin levels were independent of age (CRP: r = –0.04, p = .671; albumin: r = –0.08, p = .494), gender (CRP: Z = –1.19; p = .236; albumin: Z = –0.16; p = .872), socioeconomic status (CRP: r = 0.07, p = .557; albumin: r = 0.11, p = .427), weight (CRP: r = 0.02, p = .897; albumin: r = 0.08, p = .578), global functioning (CRP: r = 0.06, p = .555, albumin: r = 0.02, p = .884), and duration of illness (CRP: r = 0.22, p = .112; albumin: r = –0.07, p = .616).

Discussion

This study found evidence that children and adolescents with severe behavioral dysregulation, as measured by the CBCL-DP, show signs of a proinflammatory state, specifically increased CRP and decreased albumin. These associations could not be attributed to differences in age, sex, socioeconomic status, weight, global functioning, or duration of illness.

Lifestyle-related risk factors including smoking and alcohol use, higher BMI, and reduced sleep can be associated with systemic inflammation (Fuligni et al., 2009; Irwin et al., 2008). While substance use was not assessed in this study, groups did not differ in terms of weight (despite a higher number of controls diagnosed with anorexia nervosa), and no correlation was found between weight and APP levels. Sleep deprivation may have contributed to our findings, as reduced need for sleep was more prevalent among participants with the CBCL-DP than controls, and was associated with lower albumin levels. While associations between suicidality and inflammation have been reported (Lindqvist et al., 2009), APP levels were not linked to suicidality in our study. However, the CBCL probably provides rather a rough estimate of suicidality (Radeloff et al., 2012).

The high number of children with conduct disorder (CD) in the CBCL-DP group does not provide an unambiguous explanation for the altered immune function (Stadler, 2012): contrasting all subjects with CD (ICD-10 F90.1, and F91) with the remaining subjects, albumin levels were lower in CD subjects (Z = –3.20, p = .001), but there were no differences between CRP levels (Z = .76, p = .448).

Growing evidence indicates a link between bipolar disorder (BD) and inflammation (Goldstein et al., 2009). For instance, adolescents at risk for BD are more than twice as likely to express an aberrant inflammation-related messenger RNA signature compared to controls (Padmos et al., 2008). However, many researchers remain skeptic about the relationship of the CBCL-DP and BD (Holtmann et al., 2008; Holtmann, Buchmann et al., 2011; Volk & Todd, 2006), and none of our patients was diagnosed with BD.

Various studies demonstrate that early stressors, childhood adversity, and psychosocial adversity are associated with enhanced inflammatory responsiveness (Danese, Pariante, Caspi, Taylor, & Poulton, 2007; G. E. Miller & Cole, 2012; Pervanidou, 2008; Pollitt et al., 2007). In our sample, socioeconomic status was unrelated to inflammation, though stressful life events were not assessed. However, higher levels of childhood adversity have been observed in CBCL-DP children compared to psychiatric controls (Juksch et al., 2011).

A recent rapid tryptophan depletion study provided first hints of alterations in the serotonergic system in children with the CBCL-DP (Zepf et al., 2008). Given the close relationship between immune function and the serotonergic system (Miller et al., 2009), it could be speculated that the proinflammatory state is at least in part linked to the serotonergic dysfunction in CBCL-DP.

Clinical implications of our findings are not yet clear, since the cumulative effects of a low-grade inflammation on both the brain and the body remain open to speculation. In adults, CRP-values ≥ 0.30 mg/dl, reached by significantly more CBCL-DP participants than controls, are associated with high cardiovascular risk (Pearson et al., 2003). Childhood CRP values are correlated with early precursors of atherosclerosis (Järvisalo et al., 2002) and are predictive of adult CRP (Juonala et al., 2006). However, the long-term cardiovascular risks of increased CRP in childhood is uncertain, and useful cutoff points for CRP to define patients at high risk in pediatric populations are unknown (Ford et al., 2003).

There are several limitations to our findings. The primary limitation is the lack of a healthy control group. In addition, the cross-sectional study design precludes causal inferences. Our sample is drawn from the patient population of a tertiary care center, meaning we might have studied a biased sample with more severely affected patients than those in a child psychiatry outpatient practice. In addition, the sample may be biased toward a predominantly urban population. No structured interviews were used to establish ICD-10 diagnoses. Another limitation is use of a single APP measurement. Because APP concentrations show substantial intraindividual variability repeated measurements are recommended. However, to rule out an infectious etiology of elevated CRP, participants with CRP values > 1.0 mg/dl were excluded from the analysis. The main strength is the use of a highly heritable behavioral phenotype assessed by a standardized parental questionnaire with excellent psychometric properties. However, it is unclear how the findings relate to the new diagnoses of DMDD. Future research therefore should replicate these findings in a sample of children and adolescents meeting the diagnostic criteria of DMDD. Moreover, prevalence and degree of the CBCL-DP in such a clinical, well-defined sample should be compared also in light of associations to their inflammatory state.

In conclusion, our findings indicate a proinflammatory state in young patients with severe behavioral and emotional dysregulation. Besides replication in larger samples, subsequent studies should examine whether inflammatory diathesis in these children confers increased susceptibility to later development of mood disorders, cardiovascular disease and other medical morbidities. Early recognition and effective treatment of at-risk children is an important challenge for child mental health efforts (Schulte-Körne & Schiller, 2012; Schultze-Lutter, Resch, Koch, & Schimmelmann, 2011). The CBCL-DP could be useful to identify children at risk both for adverse psychopathological and physical outcome.