Elsevier

Psychoneuroendocrinology

Volume 31, Issue 10, November 2006, Pages 1245-1256
Psychoneuroendocrinology

Adrenal androgen and gonadal hormone levels in adolescent girls with conduct disorder

https://doi.org/10.1016/j.psyneuen.2006.09.005Get rights and content

Summary

There are few data on the biological correlates of female antisocial behavior. This study compared adrenal androgen and gonadal hormone levels in adolescent girls with conduct disorder (CD) to girls without any psychiatric disorder (NC). We studied 87 girls, (47 CD; 36 NC), ages 15–17 years, obtaining three blood samples, drawn 20 min apart between 8 and 9 AM in the first 72 h of the onset of menstrual flow. Plasma was assayed for testosterone, estradiol, androstenedione, dehydroepiandrosterone (DHEA), dehydroepiandrosterone-sulfate (DHEA-S), sex hormone binding globulin (SHBG), and cortisol; area under the curve (AUC) for each of the three samples was used in the data analysis. We also calculated the Free Testosterone Index, Free Estrogen Index, Index of Hyperandrogenism and cortisol to DHEA ratio. In addition to receiving a full psychiatric interview, each girl completed a self-report questionnaire on general aggression. Main hormone analyses controlled for potentially confounding variables such as psychiatric comorbidity and race. Girls with CD had significantly lower cortisol to DHEA ratios, but did not differ from NC girls on any other hormone variable. Girls with symptoms of aggressive CD had significantly higher mean free testosterone indexes, lower SHBG levels, and lower cortisol to DHEA ratios than girls with non-aggressive CD. Girls with CD scored higher on the aggression questionnaire, but there was no association between general aggression and any hormone variable for the sample. Our data suggest that girls with CD, particularly aggressive CD, have lower cortisol to DHEA ratios, higher levels of free testosterone, and lower levels of SHBG. Clinical and research implications of these findings are discussed.

Introduction

Antisocial behavior in adolescent girls is increasingly prevalent, whether defined as delinquency or conduct disorder CD, a pervasive and persistent pattern of violating the rights of others and noncompliance with rules. Between 1980 and 2003, arrest rates for assaults by girls increased 269% in the US. In 2003, girls accounted for 24% of arrests for aggravated assault, 35% of forgery arrests, and 40% of embezzlement charges for American delinquents (Synder, 2005). Epidemiologic studies show that as many as 10% of 15–17 year-old girls in the general US population meet DSM-III, DSM-III-R, or DSM-IV criteria for CD (Cohen et al., 1993; Costello et al., 2003; Kashani et al., 1989). In the UK, a population study of over 10,000 5–15 year-olds (90.8% Caucasian) reported that 3.5% of 15-year-old girls met criteria for DSM-IV criteria (5.5% of the boys the same age met criteria for CD) (Maughan et al., 2004). Two Canadian studies reported that 2.9–3.6% of 15 to 17-year-old girls met DSM-III-R criteria for CD (Breton et al., 1999; Romano et al., 2001). A large cohort study of New Zealand children followed from birth reported that 8% of the girls met DSM-IV criteria for CD at age 15 (Moffitt et al., 2001). CD is the second most common psychiatric disorder in epidemiologic studies of adolescent girls in the US, UK, and New Zealand (Cohen et al., 1987; Graham and Rutter, 1973; Kashani et al., 1987; McGee et al., 1990; Rutter et al., 1970).

Adolescent antisocial behavior exacts high societal costs and causes significant suffering to families and individuals (Jones et al., 2002; Scott et al., 2001). Furthermore, such behavior in girls is not confined to adolescence; it is associated with a 40-fold increase in adult criminality, an increased rate of early mortality, depression and substance use, poor physical health, and a high rate of intergenerational transmission of antisocial behavior (Pajer, 1998).

Unfortunately, treatment for CD in girls is not very successful. This may be partially due to our limited understanding of the mechanisms underlying female antisocial behavior. This gap is particularly evident with biological factors, such as possible effects of adrenal androgen or gonadal hormone levels. These hormones have been studied in male antisocial behavior because of a hypothesized role for androgens in the etiology of violence, but only two studies have examined this question in girls. Maras et al. (2003) measured plasma testosterone, 5-α-dihydrotestosterone (DHT) levels, and externalizing behavior in 87 14-year-old boys and girls in Germany. Groups were assigned according to behavioral scores on the Child Behavioral Checklist (CBCL) and the Teacher Report Form (TRF). There were no significant correlations between hormone levels and behavior for girls. Granger and colleagues studied 107 girls and 106 boys (mean age of 13 years) to assess the relationship between diurnal variation in salivary testosterone levels and problem behavior (Granger et al., 2003). Problem behavior was defined as high externalizing behavior scores and DSM-IV symptoms of CD or oppositional defiant disorder (ODD). A steeper decline in testosterone levels throughout the day was associated with problem behavior in girls.

Research on hormone levels in women with antisocial behavior is also meager. Dabbs et al. (Dabbs and Hargrove, 1997) measured salivary testosterone levels and behavior in 87 women, 17–60 years old, in a maximum security prison in the US. Testosterone levels were not associated with criminal violence, but were associated with aggressive-dominant behavior in prison. However, this was only the case in the younger women. His group also measured salivary testosterone concentrations in 29 young men and women (mean age was 25 years) living in an “area frequented by young people of a rebellious nature …” (Banks and Dabbs, 1996). They were compared to age- and sex-matched local college students. There was no significant correlation between the number of antisocial acts and testosterone in the women in the “delinquent” group.

It is difficult to draw any conclusions from these studies for three reasons. First, there are too few studies and the samples are not representative of the population of females with antisocial behavior. Second, testosterone was the only hormone studied; there are no data about any other hormones. Even if we focus on the broader issue of general aggression (e.g., aggression in sports), as is often done with males, aggression in females has been less often associated with testosterone than with androstenedione, DHEA, or DHEA-S levels (Albert et al., 1990, Albert et al., 1991, Albert et al., 1992; Fishbein, 1992; Inoff-Germain et al., 1988; Kapusta, 1998; Mazur, 1995; Susman et al., 1987a, Susman et al., 1987b). Third, gonadal and adrenal hormone levels are affected by variables such as menstrual cycle phase (Abraham, 1974; Bloch et al., 1998; Couchman and Hammond, 1999), body mass index, race, and social class (Ukkola et al., 2001). None of the studies controlled for these factors, making the results difficult to interpret.

Although we do not know if females with antisocial behavior are similar to males, data from males can be used to generate hypotheses about hormone—antisocial behavior associations. Testosterone levels are higher in men with antisocial behavior who are violent or socially dominant in prison (Banks and Dabbs, 1996; Dabbs et al., 1991; Ehrenkranz et al., 1974; Kreuz and Rose, 1972; Mazur, 1995). Again, testosterone is the only hormone that has been studied. Unlike men with aggressive antisocial behavior, there is no consistent association between testosterone and aggression as a personality trait in the population of Non-antisocial boys or men (Gerra, 1997; Meyer-Bahlburg et al., 1974; Rowe et al., 2004; Schaal et al., 1996).

These findings highlight the concept that general aggression and antisocial behavior are not synonymous. Indeed, aggressive behaviors are only one sub-domain of the DSM-IV criteria for CD (see Table 1) and aggression can also be part of other disorders such as intermittent explosive disorder (APA, 1994). Furthermore, general aggression is part of a healthy repertoire of self-defining behaviors, even in girls.

Studies of boys with antisocial behavior are less conclusive than those of men, although more hormones have been studied. Testosterone levels in boys with CD or delinquency have been reported as elevated (Maras et al., 2003; Scerbo and Kolko, 1994), lower (Booth et al., 2003), or not different from comparison groups (Constantino et al., 1993; Dmitrieva et al., 2001; Granger et al., 2003; Mattson et al., 1980; Rowe et al., 2004; van Goozen et al., 1998). DHEA-S levels have been reported to be elevated in boys with CD or ODD (Dmitrieva et al., 2001; van Goozen et al., 1998, van Goozen et al., 2000). Androstenedione (van Goozen et al., 1998) and estradiol (Dmitrieva et al., 2001) levels did not differ in boys with CD or ODD, but were positively correlated with high delinquency scores on behavior ratings (Susman et al., 1987a, Susman et al., 1987b). The conflicting findings in males are likely to be due to differences in sample definition, heterogeneity in pubertal stage, small sample sizes, differences in techniques for measuring hormones, and no assessment of the effect of possible confounders such as demographic variables.

In summary, there are too few studies on girls or women with antisocial behavior to draw conclusions about whether adrenal androgen or gonadal hormones are important biological correlates. However, data from boys led us to hypothesize that girls with antisocial behavior may have higher androgen levels (from either adrenal or ovarian sources) or lower estradiol levels. Therefore, we conducted this study to answer three questions. First, is adolescent antisocial behavior in girls associated with abnormal levels of adrenal androgen or gonadal hormones? Second, do girls with antisocial behavior display higher levels of general aggression than girls without such behavior? Third, is there an association between general aggression and adrenal or gonadal hormones in adolescent girls if antisocial behavior is not present?

Section snippets

Participants

We recruited 15–17 year old girls from the community, through newspaper ads and posters. The ads sought girls with behavior problems such as “truancy, fighting, stealing, and lying.” The ads also stated that we were interested in girls without these problems. Details of the recruitment and subject selection are presented elsewhere (Pajer et al., 2001a).

Exclusion criteria for both groups were: age outside the 15–17-year range, pubertal development less than Tanner Stage V, history of head trauma

Participant characteristics

The demographic and psychiatric profiles of the sample are displayed in Table 2. Girls in the CD group were slightly, but significantly older than the girls in the NC group. Ethnicity and social classes proportions were not significantly different. None of the NC girls had any psychiatric illness by design. The three most common co-morbid psychiatric disorders in the CD sub-sample were some type of substance use disorder (38%), anxiety disorders (25.3%), and depressive disorders (10.6%). Two

Discussion

This study improved upon methodologic problems in previous studies by using girls with research diagnoses of CD and comparing them to girls without any psychiatric or medical disorders. Our sample was homogeneous with regards to pubertal stage. The data collection occurred at a standardized time (during the day and the menstrual cycle), we collected three samples, and we investigated more hormones than other studies, also assessing free levels of testosterone and estradiol. Our data analysis

Acknowledgments

This study was supported by a grant from NIMH (MH01285-04, PI: Pajer). The authors are grateful for the efforts of Judith Navratil, Barbara LeBeau, Shirley Conti, Stephen Neal, Deborah McCaskill, and Phyllis Wanetick in collecting these data and for the additional assistance of James Perel, Ph.D. and his laboratory. The authors also thank the young women who participated in this study.

References (86)

  • A. Maras et al.

    Association of testosterone and dihydrotestosterone with externalizing behavior in adolescent boys and girls

    Psychoneuroendocrinology

    (2003)
  • C.A. Martin et al.

    The association of testosterone with nicotine use in young adult females

    Addict. Behav.

    (2001)
  • A. Mazur

    Biosocial models of deviant behavior among male army veterans

    Biol. Psychiatry

    (1995)
  • R. McGee et al.

    DSM-III disorders in a large sample of adolescents

    J. Am. Acad. Child Adolesc. Psychol

    (1990)
  • J.C. Pruessner et al.

    Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change

    Psychoneuroendocrinology

    (2003)
  • R. Rowe et al.

    Testosterone, antisocial behavior, and social dominance in boys: pubertal development and biosocial interaction

    Biol. Psychiatry

    (2004)
  • A.S. Scerbo et al.

    Salivary testosterone and cortisol in disruptive children: relationship to aggressive, hyperactive, and internalizing behaviors

    J. Am. Acad. Child Adoles. Psychiatry

    (1994)
  • B. Schaal et al.

    Male testosterone linked to high social dominance but low physical aggression in early adolescence

    J. Am. Acad. Child Adolesc. Psychiatry

    (1996)
  • C. Sundblad et al.

    Reduced extracellular levels of serotonin in the amygdala of androgenized female rats

    Eur. Neuropsychopharmacol.

    (1997)
  • E. Susman et al.

    Hormonal influences on aspects of psychological development during adolescence

    J. Adoles. Health Care

    (1987)
  • S.H. van Goozen et al.

    Adrenal androgens and aggression in conduct disorder prepubertal boys and normal controls

    Biol. Psychiatry

    (1998)
  • S.H. van Goozen et al.

    Increased adrenal androgen functioning in children with oppositional defiant disorder: a comparison with psychiatric and normal controls

    J. Am. Acad. Child Adolesc. Psychiatry

    (2000)
  • I. Wiegratz et al.

    Effect of two oral contraceptives containing ethinylestradiol and gestodene or norgestimate upon androgen parameters and serum binding proteins

    Contraception

    (1995)
  • G.E. Abraham

    Ovarian and adrenal contribution to peripheral androgens during the menstrual cycle

    J. Clin. Endocrinol. Metab.

    (1974)
  • A. Aluja et al.

    Hostility-aggressiveness, sensation seeking, and sex hormones in men: re-exploring their relationship

    Neuropsychobiology

    (2004)
  • A. Angold et al.

    Pubertal changes in hormone levels and depression in girls

    Psychol. Med.

    (1999)
  • APA. (1994). Diagnostic and Statistical Manual of Mental Disorders, fourth ed. (DSM-IV). Washington, DC: American...
  • T. Banks et al.

    Salivary testosterone and cortisol in a delinquent and violent urban subculture

    J. Social Psychol.

    (1996)
  • K. Bjorkqvist et al.

    Do girls manipulate and boys fight? Developmental trends in regard to direct and indirect aggression

    Aggressive Behav.

    (1992)
  • A. Booth et al.

    Testosterone and child and adolescent adjustment: the moderating role of parent–child relationships

    Dev. Psychol.

    (2003)
  • J.J. Breton et al.

    Quebec child mental health survey: prevalence of DSM-III-R mental health disorders

    J. Child. Psychol. Psychiatry

    (1999)
  • J. Brooks-Gunn et al.

    Validity of self-report measures of girls’ pubertal status

    Child Dev.

    (1987)
  • T.T. Brown et al.

    Endocrine effects of marijuana

    J. Clin. Pharmacol.

    (2002)
  • H.G. Burger

    Androgen production in women

    Fertil. Steril.

    (2002)
  • A.H. Buss et al.

    An inventory for assessing different kinds of hostility

    J. Consult. Psychol.

    (1957)
  • A.H. Buss et al.

    The Aggression Questionnaire

    J. Personality Social Psychol.

    (1992)
  • L. Buydens-Branchey et al.

    Cocaine addicts with conduct disorder are typified by decreased cortisol responsivity and high plasma levels of DHEA-S

    Neuropsychobiology

    (2004)
  • K. Carlstrom et al.

    Diurnal rhythm and effects of oral contraceptives on serum dehydroepiandrosterone sulfate (DHEAS) are related to alterations in serum albumin rather than to changes in adrenocortical steroid secretion

    Scand. J. Clin. Lab. Invest.

    (2002)
  • CDC. (2006). YRBSS: Youth Risk Behavior Surveillance System. Data & Statistics. Comprehensive Results: Center for...
  • M.S. Cepeda et al.

    Comparison of logistic regression versus propensity score when the number of events is low and there are multiple confounders

    Am. J. Epidemiol.

    (2003)
  • D. Cibula et al.

    The best correlation of the new index of hyperandrogenism with the grade of increased body hair

    Eur. J. Endocrinol.

    (2000)
  • P. Cohen et al.

    An epidemiological study of disorders in late childhood and adolescence: I. age and gender -specific prevalence

    J. Child Psychol. Psychiatry Allied Discip.

    (1993)
  • E.J. Costello et al.

    Prevalence and development of psychiatric disorders in childhood and adolescence

    Arch. Gen. Psychiatry

    (2003)
  • Cited by (95)

    • Dehydroepiandrosterone sulfate, free testosterone, and sex hormone-binding globulin on susceptibility to attention-deficit/hyperactivity disorder

      2019, Psychoneuroendocrinology
      Citation Excerpt :

      Free testosterone levels did not differ significantly between the ADHD and control groups, and testosterone was not significantly involved in the behavioral symptoms or neuropsychological effects of ADHD. Some researchers have identified a significant positive correlation between testosterone levels and a tendency toward disruptive behavior disorders (Pajer et al., 2006; Scerbo and Kolko, 1994). Testicular hormones are generally considered to provide resiliency against chronic stress in males (Wainwright et al., 2011).

    • Biomarkers in aggression

      2019, Advances in Clinical Chemistry
    • Serum levels of cortisol, dehydroepiandrosterone, and oxytocin in children with attention-deficit/hyperactivity disorder combined presentation with and without comorbid conduct disorder

      2018, Psychiatry Research
      Citation Excerpt :

      To our knowledge, only three previous studies have evaluated the association between conduct disorder and DHEA, all of which failed to find a relationship (Pajer et al., 2006; Dorn et al., 2009; Golubchik et al., 2009). Pajer et al. (2006) examined DHEA levels in girls 15–17 years of age with conduct disorder and failed to demonstrate any relationship between DHEA levels and conduct disorder. Golubchik et al. (2009) examined neurosteroid blood levels in adolescents with conduct disorder as compared to normal controls and showed that DHEA levels did not differ between the two groups.

    View all citing articles on Scopus
    View full text