538655
research-article2014
JADXXX10.1177/1087054714538655Journal of Attention DisordersBendiksen et al.
Article
Co-Occurrence of ODD and CD in Preschool Children With Symptoms of ADHD
Journal of Attention Disorders 1–12 © 2014 SAGE Publications Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1087054714538655 jad.sagepub.com
Bothild Bendiksen1,2, Elisabeth Svensson3, Heidi Aase4, Ted Reichborn-Kjennerud2,4, Svein Friis1,2, Anne M. Myhre1,2, and Pål Zeiner1
Abstract Objective: Patterns of co-occurrence between ADHD, Oppositional Defiant Disorder (ODD), and Conduct Disorder (CD) were examined in a sample of non-referred preschool children. ADHD subtypes and sex differences were also explored. Method: Children aged 3.5 years (n = 1,048) with high scores on ADHD characteristics were recruited from the Norwegian Mother and Child Cohort Study and clinically assessed, including a semi-structured psychiatric interview. Results: In children with ADHD, concurrent ODD was present more often than CD (31% vs. 10%), but having ADHD gave higher increase in the odds of CD than of ODD (ODD: odds ratio [OR] = 6.7, 95% confidence interval [CI] = [4.2, 10.8]; CD: OR = 17.6, 95% CI = [5.9, 52.9]). We found a greater proportion of children having the combined ADHD subtype as well as more severe inattentiveness among children with co-occurring CD compared with ODD. Sex differences were minor. Conclusion: There are important differences in co-occurring patterns of ODD and CD in preschool children with ADHD. (J. of Att. Dis. XXXX; XX(X) XX-XX) Keywords ADHD, preschool children, subtypes, sex differences, co-occurring symptoms, Oppositional Defiant Disorder, Conduct Disorder ADHD is one of the most common mental disorders affecting about 5% of children and adolescents worldwide (Polanczyk, de Lima, Horta, Biederman, & Rohde, 2007). The disorder is associated with increasing rates of comorbidity over the age span (Taurines et al., 2010). Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) are the most commonly reported co-occurring disorders in school-aged children with ADHD with rates at about 50% and 20%, respectively (Pliszka, 1998). The prevalence of comorbid ODD or CD in clinical preschool samples is similar to those in school-aged children, namely, 50% to 65% with co-occurring ODD and about 20% with co-occurring CD (Kadesjo, Hagglof, Kadesjo, & Gillberg, 2003; Posner et al., 2007; Wilens et al., 2002). In community-based preschool samples the prevalence is lower, with about 20% of children with ADHD having co-occurring ODD and 14% CD (Wichstrom et al., 2011). Although comorbidity of behavior disorders, such as ODD and CD, is reported to be associated with increased severity of ADHD symptoms in school-aged children (Connor & Ford, 2012), this has not been fully investigated in preschool children. The diagnostic system Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association [APA], 2000) defines
three ADHD subtypes: the inattentive subtype (ADHD-IA), the hyperactive–impulsive subtype (ADHD-HI), and the combined subtype (ADHD-C). In school-aged children, the most common subtypes diagnosed are ADHD-C (~60%) and ADHD-IA (~30%; Ford, Goodman, & Meltzer, 2003). In 4-year-old preschool children with ADHD, however, the most common subtypes are ADHD-HI (~60%) and ADHD-C (~30%; Byrne, Bawden, Beattie, & DeWolfe, 2000; Gadow, Sprafkin, & Nolan, 2001; Gimpel & Kuhn, 2000; Lahey et al., 1998; Wichstrom et al., 2011). In school-aged children with ADHD, the co-occurrence patterns vary between subtypes. ODD and CD show a stronger association with ADHD-HI and ADHD-C than ADHD-IA in both clinical and community samples (Eiraldi, Power, & Nezu, 1997; 1
Oslo University Hospital, Norway University of Oslo, Norway 3 University of Aarhus, Denmark 4 Norwegian Institute of Public Health, Oslo, Norway 2
Corresponding Author: Bothild Bendiksen, Department of Psychiatric Research and Development, Oslo University Hospital, University of Oslo, P.O. Box 4959 Nydalen, 0424 Oslo, Norway. Email: [email protected]
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
2
Journal of Attention Disorders
Willcutt, Pennington, Chhabildas, Friedman, & Alexander, 1999). Some preschool studies have reported a stronger association between ADHD-C and ODD (Kadesjo et al., 2003; Lavigne, Lebailly, Hopkins, Gouze, & Binns, 2009; Nolan, Gadow, & Sprafkin, 2001; Riley et al., 2008), whereas studies addressing relationships between ADHD subtypes and CD in preschoolers are lacking. A diagnosis of ADHD requires attaining a symptom level above a set cut off values. However, symptoms of hyperactivity, impulsivity, and inattention are not discrete, distinct entities, and a diagnosis may be seen more or less as an entity above a certain cut-off at one extreme on a continuum of severity (Sonuga-Barke, Auerbach, Campbell, Daley, & Thompson, 2005). Thus, a dimensional approach may be an important supplement to the diagnostic categories for the understanding of early patterns of disorders. In ADHD, there is a male predominance in older children and adolescents, and patterns of comorbidity vary between boys and girls at different ages (Monuteaux, Mick, Faraone, & Biederman, 2010). In primary-school children, boys clearly outnumber girls in ADHD-C, but there are only minor sex differences for ADHD-HI and ADHD-IA (Decker, McIntosh, Kelly, Nicholls, & Dean, 2001). In preschool, the male predominance in ADHD is less pronounced, but some studies report that boys have ADHD-C more often than girls (Egger, Kondo, & Angold, 2006; Gimpel & Kuhn, 2000), whereas some studies from preschool community samples have reported that ADHD and co-occurring behavior disorders are equally common in boys and girls (Nolan et al., 2001; Tandon, Si, & Luby, 2011), others find similar patterns to those observed in older children (Gadow et al., 2001; Wichstrom et al., 2011). This suggests some uncertainty concerning early sex-specific characteristics in the development of ADHD and co-occurring disruptive behaviors. It is plausible that boys with ADHD and co-occurring behavior problems are referred to treatment more often than girls; thus, a sex distribution toward boys in clinical samples may be influenced by referral-biases (Novik et al., 2006). Non-referred study-populations may overcome this problem and should thus be better suited to study sex distribution and differences in ADHD and comorbid disruptive behavior disorders. The number of studies on early co-occurring patterns of ODD and CD in young children with ADHD is still limited as longitudinal studies often begin in late preschool or early school age (Keenan, Shaw, Walsh, & Delliquadri, 1997). Several preschool studies on this issue have relatively small sample sizes (Keenan et al., 1997; Pierce, Ewing, & Campbell, 1999; Tandon et al., 2011). In addition, some studies consider only co-occurring ODD and not CD, or collapse the two into one disruptive behavior disorder (Bufferd, Dougherty, Carlson, Rose, & Klein, 2012; DuPaul, McGoey, Eckert, & VanBrakle, 2001; Kadesjo et al., 2003; Lavigne et al., 2009; Posner et al., 2007). Finally, some
studies have only used checklists and not diagnostic interviews for symptom measures (Gadow & Nolan, 2002; Kadesjo et al., 2003; Lavigne et al., 2009; Nolan et al., 2001). These issues indicate that studies examining preschool children with ADHD and the comorbidity patterns with ODD and CD with recommended clinical instruments are warranted. The present study aimed to investigate the co-occurrence of ODD and CD in 3-year-old non-referred children with clinical symptoms of ADHD. More specifically, we aimed to answer the following questions: What is the prevalence of co-occurring ODD and/or CD in children with clinical symptoms of ADHD? What is the risk of ODD or CD in children with ADHD? What are the associations between the numbers of ODD and CD symptoms to the number of ADHD symptoms and to what extent do ODD and CD predict the variance in symptoms of ADHD? Are comorbidity patterns of ODD or CD related to specific subtypes of ADHD? Is the severity of IA and HI symptoms related to certain patterns of co-occurring ODD or CD? Are there sex differences in the prevalence of ADHD, its subtypes, or in the co-occurring patterns of ODD and CD in non-referred young preschoolers?
Method Participants All participants were recruited from the Norwegian Mother and Child Cohort Study (MoBa), a populationbased prospective birth cohort study of about 107,000 children run by the Norwegian Institute of Public Health (Magnus, Haug, Nystad, & Skjaerven , 2006). To identify a large number of preschoolers who might be at risk of developing ADHD, 3-year-old children whose sum score on 11 questions regarding hyperactivity, impulsivity, and inattention in the 36 months MoBa questionnaire, was above the 90th percentile, were invited to a clinical assessment. Six of these 11 questions were selected from the Child Behavior Checklist (Achenbach, 2000) and 5 questions were from the symptom criteria for ADHD in DSMIV-TR (APA, 2000). Of the 2,798 invited children 1,048 (37.5%) completed the clinical assessments including diagnostic evaluation. There were no statistically significant differences between participants and invited non-participants regarding background factors and pre-and perinatal risk factors except for a higher level of maternal education in participants. Figure 1 shows enrollment into the present study. The children, when aged 36 to 44 months, participated in a 1-day clinical assessment at Oslo University Hospital together with at least one parent. The exclusion criteria were severe medical conditions or high scores on autistic symptoms. One of the parents had to speak Norwegian language.
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
3
Bendiksen et al.
Total Norwegian Mother and child cohort (MoBa). Enrollment: 1999-2008 MoBa sample n = 107.000. Participation rate: 38.7%
Eligible children within sampling frame (Born between April 2004 and January 2008). Available questionnaires at 36 months within sampling frame: n = 33050.
Invited to clinical assessment at 36 months ”ADHD Risk”: children with high scores on ADHD characteristics: n = 2798
Participants in clinical assess ment at mean age 42 months ”ADHD Risk”: n = 1048 (37.5% of invited)
ADHD
No ADHD
n = 163 (15.5%)
n = 885 (84.5%)
ODD
CD
ODD + CD
ODD
CD
ODD + CD
n = 45
n = 12
n=5
n = 51
n=5
n=8
Figure 1. Enrollment of children into the “preschool ADHD study.”
Note. ADHD = DSM-IV-TR diagnosis of ADHS; ODD = DSM-IV-TR diagnosis of Oppositional Defiant Disorder; CD = DSM-IV-TR diagnosis of Conduct Disorder.
Measures One of the parents, most often the mother, was interviewed with “The Preschool Age Psychiatric Assessment” (PAPA; Egger & Angold, 2004). This semi-structured interview provides information about psychiatric symptoms, including frequency, intensity, age of onset, and presence in different settings relevant for preschool children. The task of the interviewer is to ensure that the interviewee understands the questions and that she or he provides clear information concerning the symptom at hand. Interviewers continue to probe until there is enough information to decide whether the symptom is present at pre-specified levels of severity. If so,
its onset date is recorded along with its frequency of occurrence and its presence at home and at day care. A 3-month primary period is used as the behavior of preschool children change rapidly in this period (Egger & Angold, 2004). Diagnoses were generated using algorithms implementing criteria from the DSM-IV-TR (APA, 2000). According to DSM-IV-TR diagnostic criteria a diagnosis of ADHD requires (criterion A) at least six (of nine) inattentive (ADHD-IA subtype), or six (of nine) hyperactive–impulsive symptoms (ADHD-HI subtype). If both subtypes are present, criteria for the combined subtype (ADHD-C) are met. In addition, symptoms must be maladaptive and inconsistent with developmental level and cause impairment.
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
4
Journal of Attention Disorders
At the end of each chapter in PAPA, six functional domains provide information of impairment/impact. Impairment was scored as present when symptoms caused reduced function in two or more of the following functional areas; “relationship with parents and the rest of the family,” “the ability to keep friends,” “academic performance,” “play and leisure activities,” “quality of life,” or “as a burden to the family”(Egger & Angold, 2004; Goodman, 1997). Each subscale of impairment ranged from 0 to 3 and the total impairment score from 0 to 18. Severity of ADHD was assessed by total symptom counts of HI, IA, and total ADHD symptoms (Connor & Ford, 2012). Each symptom was scored 0 (no), 2 (yes, moderate), 3 (yes, severe). Correlations between the number of symptoms and severity scores were .97 for IA symptoms, .97 for HI symptoms, and .98 for ADHD symptoms, thus, measures of total symptom counts were used as they equated measures of symptom severity. The DSM-IV-TR diagnostic criteria of ODD and CD include 8 and 15 diagnostic symptoms, respectively. In PAPA, symptoms of ODD are modified by cutpoints based on the top 10% of frequency for preschoolers (Egger et al., 2006). Five diagnostic symptoms for CD were excluded because they are not applicable to preschoolers (“stealing with confrontation,” “forced sexual activity,” “breaking into a house or car,” “running away from home,” and “truancy”). The modified and more age-appropriate criteria include eight symptoms, and have shown moderate validity (Keenan et al., 2007; Egger & Angold, 2004). A DSMIV-TR diagnosis of ODD requires four or more oppositional symptoms and a diagnosis of CD three or more conduct symptoms, respectively. According to PAPA, symptoms must be present for at least 3 months and cause functional impairment. The PAPA interviews were performed by trained psychology students and supervised by a clinically trained psychologist or child psychiatrists. An inter-rater reliability check was carried out by a second rater, blind to any knowledge about the child and family, and rescored from audiotapes of 79 randomly selected interviews. The average intraclass correlations (ICC) were .98 for total number of ADHD symptoms, .94 for total impairment score of ADHD, .98 for total number of ODD symptoms, .85 for impairment score of ODD, .91 for total number of CD symptoms, and .99 for impairment score of CD. Covariates. Covariates included maternal and paternal educational level (socioeconomic status [SES]) obtained from the MoBa questionnaire at 17th gestational week. Information about maternal age at delivery and marital status were obtained from The Norwegian Medical Birth Registry (Magnus, Haug, Nystad, & Skjaerven, 2006). From the neuropsychological assessment performed by a trained psychologist, a short form of Stanford Binet 5th edition provided
information about general intellectual functioning (Roid, 2003). The Regional Ethics Committee and the Norwegian data inspectorate approved the study in 2007. Assessments were carried out according to ethical standards, and principles of the Helsinki declaration were followed. Parents returned a written consent prior to the clinical assessment.
Data Analysis Maternal, paternal, and offspring characteristics for all participants were compared according to presence of ADHD (yes/no) in the child. We examined statistical differences in background characteristic by independent-sample t tests for continuous measures and Pearson chi-square tests or Fisher’s exact tests for categorical measures. Covariates significantly associated with ADHD were adjusted for in the subsequent multiple regression analyses. We examined the odds ratios (ORs) of having co-occurring ODD and/or CD in ADHD by binary logistic regression and adjusted for possible confounding by gender, child’s age (months), IQ, maternal age (years), marital status (cohabiting/married or single parent), maternal education (years), and paternal education (years). We examined for possible confounding or interactions between independent variables (i.e., sex, ODD, and CD). The associations between the numbers of ODD and/or CD symptoms to the numbers of ADHD symptoms were examined by Pearson product–moment correlations and by hierarchical multiple linear regression analyses. There were no statistically significant non-linear relationships of ODD and CD symptoms, nor were there significant interactions between independent variables (i.e., sex, ODD, and CD symptoms) in predicting the number of ADHD symptoms. Comorbidity patterns of ODD and/or CD across different subtypes were compared by crosstabs and Mantel Haenzel’s ORs were reported. Furthermore, a one-way ANOVA between the groups was conducted to examine the relationship between comorbidity of ODD or CD and the severity of IA and HI symptoms. Participants were divided into four groups: ADHD only, ADHD plus ODD, ADHD plus CD, and ADHD plus ODD plus CD. Post hoc comparisons with Bonferroni correction were used. Finally, we examined sex distribution and sex differences in ADHD, subtypes of ADHD, and in co-occurring symptom clusters by independent-sample t tests for continuous measures and Pearson chi-square tests or Fisher’s exact tests for categorical measures. All tests were two-tailed.
Results Prevalence rate of children fulfilling the DSM-IV-TR diagnostic criteria for ADHD was 16% (n = 163). Table 1 shows
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
5
Bendiksen et al. Table 1. Sample Characteristics of 1,048 Preschool Children in the ADHD Study.
ADHD (n = 163)
No ADHD (n = 885)
n (%)
n (%)
97 (59.5) Sex (boys) Age (months) IQ ADHD symptoms IA symptoms HI symptoms ODD symptoms CD symptoms Mothers educational level (years) Fathers educational level (years) Mothers age (years)
448 (50.6)
M (SD)
M (SD)
41.6 (1.3) 99.5 (9.0) 10.6 (2.8) 3.8 (2.4) 6.8 (1.3) 2.8 (1.9) 1.3 (1.3) 14.4 (2.5) 13.6 (2.9) 29.9 (4.4)
41.8 (1.3) 102.2 (9.2) 3.1 (2.7) 1.0 (1.3) 2.2 (1.9) 1.5 (1.4) 0.6 (0.9) 15.4 (2.3) 14.6 (2.7) 30.6 (4.2)
n (%)
Civil status (single parent)
9 (5.5)
n (%) 27 (3.1)
χ2/t test χ2(df)
p
4.4 (1)
.04
t(df)
p
−1.3 (1,044) −3.3 (1,033) 30.5 (1,046) 14.3 (181) 38.9 (322) 8.3 (198) 6.8 (186) −4.4 (209) −3.9 (205) −2.1 (1,045)
.19 .001
research-article2014
JADXXX10.1177/1087054714538655Journal of Attention DisordersBendiksen et al.
Article
Co-Occurrence of ODD and CD in Preschool Children With Symptoms of ADHD
Journal of Attention Disorders 1–12 © 2014 SAGE Publications Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1087054714538655 jad.sagepub.com
Bothild Bendiksen1,2, Elisabeth Svensson3, Heidi Aase4, Ted Reichborn-Kjennerud2,4, Svein Friis1,2, Anne M. Myhre1,2, and Pål Zeiner1
Abstract Objective: Patterns of co-occurrence between ADHD, Oppositional Defiant Disorder (ODD), and Conduct Disorder (CD) were examined in a sample of non-referred preschool children. ADHD subtypes and sex differences were also explored. Method: Children aged 3.5 years (n = 1,048) with high scores on ADHD characteristics were recruited from the Norwegian Mother and Child Cohort Study and clinically assessed, including a semi-structured psychiatric interview. Results: In children with ADHD, concurrent ODD was present more often than CD (31% vs. 10%), but having ADHD gave higher increase in the odds of CD than of ODD (ODD: odds ratio [OR] = 6.7, 95% confidence interval [CI] = [4.2, 10.8]; CD: OR = 17.6, 95% CI = [5.9, 52.9]). We found a greater proportion of children having the combined ADHD subtype as well as more severe inattentiveness among children with co-occurring CD compared with ODD. Sex differences were minor. Conclusion: There are important differences in co-occurring patterns of ODD and CD in preschool children with ADHD. (J. of Att. Dis. XXXX; XX(X) XX-XX) Keywords ADHD, preschool children, subtypes, sex differences, co-occurring symptoms, Oppositional Defiant Disorder, Conduct Disorder ADHD is one of the most common mental disorders affecting about 5% of children and adolescents worldwide (Polanczyk, de Lima, Horta, Biederman, & Rohde, 2007). The disorder is associated with increasing rates of comorbidity over the age span (Taurines et al., 2010). Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) are the most commonly reported co-occurring disorders in school-aged children with ADHD with rates at about 50% and 20%, respectively (Pliszka, 1998). The prevalence of comorbid ODD or CD in clinical preschool samples is similar to those in school-aged children, namely, 50% to 65% with co-occurring ODD and about 20% with co-occurring CD (Kadesjo, Hagglof, Kadesjo, & Gillberg, 2003; Posner et al., 2007; Wilens et al., 2002). In community-based preschool samples the prevalence is lower, with about 20% of children with ADHD having co-occurring ODD and 14% CD (Wichstrom et al., 2011). Although comorbidity of behavior disorders, such as ODD and CD, is reported to be associated with increased severity of ADHD symptoms in school-aged children (Connor & Ford, 2012), this has not been fully investigated in preschool children. The diagnostic system Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association [APA], 2000) defines
three ADHD subtypes: the inattentive subtype (ADHD-IA), the hyperactive–impulsive subtype (ADHD-HI), and the combined subtype (ADHD-C). In school-aged children, the most common subtypes diagnosed are ADHD-C (~60%) and ADHD-IA (~30%; Ford, Goodman, & Meltzer, 2003). In 4-year-old preschool children with ADHD, however, the most common subtypes are ADHD-HI (~60%) and ADHD-C (~30%; Byrne, Bawden, Beattie, & DeWolfe, 2000; Gadow, Sprafkin, & Nolan, 2001; Gimpel & Kuhn, 2000; Lahey et al., 1998; Wichstrom et al., 2011). In school-aged children with ADHD, the co-occurrence patterns vary between subtypes. ODD and CD show a stronger association with ADHD-HI and ADHD-C than ADHD-IA in both clinical and community samples (Eiraldi, Power, & Nezu, 1997; 1
Oslo University Hospital, Norway University of Oslo, Norway 3 University of Aarhus, Denmark 4 Norwegian Institute of Public Health, Oslo, Norway 2
Corresponding Author: Bothild Bendiksen, Department of Psychiatric Research and Development, Oslo University Hospital, University of Oslo, P.O. Box 4959 Nydalen, 0424 Oslo, Norway. Email: [email protected]
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
2
Journal of Attention Disorders
Willcutt, Pennington, Chhabildas, Friedman, & Alexander, 1999). Some preschool studies have reported a stronger association between ADHD-C and ODD (Kadesjo et al., 2003; Lavigne, Lebailly, Hopkins, Gouze, & Binns, 2009; Nolan, Gadow, & Sprafkin, 2001; Riley et al., 2008), whereas studies addressing relationships between ADHD subtypes and CD in preschoolers are lacking. A diagnosis of ADHD requires attaining a symptom level above a set cut off values. However, symptoms of hyperactivity, impulsivity, and inattention are not discrete, distinct entities, and a diagnosis may be seen more or less as an entity above a certain cut-off at one extreme on a continuum of severity (Sonuga-Barke, Auerbach, Campbell, Daley, & Thompson, 2005). Thus, a dimensional approach may be an important supplement to the diagnostic categories for the understanding of early patterns of disorders. In ADHD, there is a male predominance in older children and adolescents, and patterns of comorbidity vary between boys and girls at different ages (Monuteaux, Mick, Faraone, & Biederman, 2010). In primary-school children, boys clearly outnumber girls in ADHD-C, but there are only minor sex differences for ADHD-HI and ADHD-IA (Decker, McIntosh, Kelly, Nicholls, & Dean, 2001). In preschool, the male predominance in ADHD is less pronounced, but some studies report that boys have ADHD-C more often than girls (Egger, Kondo, & Angold, 2006; Gimpel & Kuhn, 2000), whereas some studies from preschool community samples have reported that ADHD and co-occurring behavior disorders are equally common in boys and girls (Nolan et al., 2001; Tandon, Si, & Luby, 2011), others find similar patterns to those observed in older children (Gadow et al., 2001; Wichstrom et al., 2011). This suggests some uncertainty concerning early sex-specific characteristics in the development of ADHD and co-occurring disruptive behaviors. It is plausible that boys with ADHD and co-occurring behavior problems are referred to treatment more often than girls; thus, a sex distribution toward boys in clinical samples may be influenced by referral-biases (Novik et al., 2006). Non-referred study-populations may overcome this problem and should thus be better suited to study sex distribution and differences in ADHD and comorbid disruptive behavior disorders. The number of studies on early co-occurring patterns of ODD and CD in young children with ADHD is still limited as longitudinal studies often begin in late preschool or early school age (Keenan, Shaw, Walsh, & Delliquadri, 1997). Several preschool studies on this issue have relatively small sample sizes (Keenan et al., 1997; Pierce, Ewing, & Campbell, 1999; Tandon et al., 2011). In addition, some studies consider only co-occurring ODD and not CD, or collapse the two into one disruptive behavior disorder (Bufferd, Dougherty, Carlson, Rose, & Klein, 2012; DuPaul, McGoey, Eckert, & VanBrakle, 2001; Kadesjo et al., 2003; Lavigne et al., 2009; Posner et al., 2007). Finally, some
studies have only used checklists and not diagnostic interviews for symptom measures (Gadow & Nolan, 2002; Kadesjo et al., 2003; Lavigne et al., 2009; Nolan et al., 2001). These issues indicate that studies examining preschool children with ADHD and the comorbidity patterns with ODD and CD with recommended clinical instruments are warranted. The present study aimed to investigate the co-occurrence of ODD and CD in 3-year-old non-referred children with clinical symptoms of ADHD. More specifically, we aimed to answer the following questions: What is the prevalence of co-occurring ODD and/or CD in children with clinical symptoms of ADHD? What is the risk of ODD or CD in children with ADHD? What are the associations between the numbers of ODD and CD symptoms to the number of ADHD symptoms and to what extent do ODD and CD predict the variance in symptoms of ADHD? Are comorbidity patterns of ODD or CD related to specific subtypes of ADHD? Is the severity of IA and HI symptoms related to certain patterns of co-occurring ODD or CD? Are there sex differences in the prevalence of ADHD, its subtypes, or in the co-occurring patterns of ODD and CD in non-referred young preschoolers?
Method Participants All participants were recruited from the Norwegian Mother and Child Cohort Study (MoBa), a populationbased prospective birth cohort study of about 107,000 children run by the Norwegian Institute of Public Health (Magnus, Haug, Nystad, & Skjaerven , 2006). To identify a large number of preschoolers who might be at risk of developing ADHD, 3-year-old children whose sum score on 11 questions regarding hyperactivity, impulsivity, and inattention in the 36 months MoBa questionnaire, was above the 90th percentile, were invited to a clinical assessment. Six of these 11 questions were selected from the Child Behavior Checklist (Achenbach, 2000) and 5 questions were from the symptom criteria for ADHD in DSMIV-TR (APA, 2000). Of the 2,798 invited children 1,048 (37.5%) completed the clinical assessments including diagnostic evaluation. There were no statistically significant differences between participants and invited non-participants regarding background factors and pre-and perinatal risk factors except for a higher level of maternal education in participants. Figure 1 shows enrollment into the present study. The children, when aged 36 to 44 months, participated in a 1-day clinical assessment at Oslo University Hospital together with at least one parent. The exclusion criteria were severe medical conditions or high scores on autistic symptoms. One of the parents had to speak Norwegian language.
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
3
Bendiksen et al.
Total Norwegian Mother and child cohort (MoBa). Enrollment: 1999-2008 MoBa sample n = 107.000. Participation rate: 38.7%
Eligible children within sampling frame (Born between April 2004 and January 2008). Available questionnaires at 36 months within sampling frame: n = 33050.
Invited to clinical assessment at 36 months ”ADHD Risk”: children with high scores on ADHD characteristics: n = 2798
Participants in clinical assess ment at mean age 42 months ”ADHD Risk”: n = 1048 (37.5% of invited)
ADHD
No ADHD
n = 163 (15.5%)
n = 885 (84.5%)
ODD
CD
ODD + CD
ODD
CD
ODD + CD
n = 45
n = 12
n=5
n = 51
n=5
n=8
Figure 1. Enrollment of children into the “preschool ADHD study.”
Note. ADHD = DSM-IV-TR diagnosis of ADHS; ODD = DSM-IV-TR diagnosis of Oppositional Defiant Disorder; CD = DSM-IV-TR diagnosis of Conduct Disorder.
Measures One of the parents, most often the mother, was interviewed with “The Preschool Age Psychiatric Assessment” (PAPA; Egger & Angold, 2004). This semi-structured interview provides information about psychiatric symptoms, including frequency, intensity, age of onset, and presence in different settings relevant for preschool children. The task of the interviewer is to ensure that the interviewee understands the questions and that she or he provides clear information concerning the symptom at hand. Interviewers continue to probe until there is enough information to decide whether the symptom is present at pre-specified levels of severity. If so,
its onset date is recorded along with its frequency of occurrence and its presence at home and at day care. A 3-month primary period is used as the behavior of preschool children change rapidly in this period (Egger & Angold, 2004). Diagnoses were generated using algorithms implementing criteria from the DSM-IV-TR (APA, 2000). According to DSM-IV-TR diagnostic criteria a diagnosis of ADHD requires (criterion A) at least six (of nine) inattentive (ADHD-IA subtype), or six (of nine) hyperactive–impulsive symptoms (ADHD-HI subtype). If both subtypes are present, criteria for the combined subtype (ADHD-C) are met. In addition, symptoms must be maladaptive and inconsistent with developmental level and cause impairment.
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
4
Journal of Attention Disorders
At the end of each chapter in PAPA, six functional domains provide information of impairment/impact. Impairment was scored as present when symptoms caused reduced function in two or more of the following functional areas; “relationship with parents and the rest of the family,” “the ability to keep friends,” “academic performance,” “play and leisure activities,” “quality of life,” or “as a burden to the family”(Egger & Angold, 2004; Goodman, 1997). Each subscale of impairment ranged from 0 to 3 and the total impairment score from 0 to 18. Severity of ADHD was assessed by total symptom counts of HI, IA, and total ADHD symptoms (Connor & Ford, 2012). Each symptom was scored 0 (no), 2 (yes, moderate), 3 (yes, severe). Correlations between the number of symptoms and severity scores were .97 for IA symptoms, .97 for HI symptoms, and .98 for ADHD symptoms, thus, measures of total symptom counts were used as they equated measures of symptom severity. The DSM-IV-TR diagnostic criteria of ODD and CD include 8 and 15 diagnostic symptoms, respectively. In PAPA, symptoms of ODD are modified by cutpoints based on the top 10% of frequency for preschoolers (Egger et al., 2006). Five diagnostic symptoms for CD were excluded because they are not applicable to preschoolers (“stealing with confrontation,” “forced sexual activity,” “breaking into a house or car,” “running away from home,” and “truancy”). The modified and more age-appropriate criteria include eight symptoms, and have shown moderate validity (Keenan et al., 2007; Egger & Angold, 2004). A DSMIV-TR diagnosis of ODD requires four or more oppositional symptoms and a diagnosis of CD three or more conduct symptoms, respectively. According to PAPA, symptoms must be present for at least 3 months and cause functional impairment. The PAPA interviews were performed by trained psychology students and supervised by a clinically trained psychologist or child psychiatrists. An inter-rater reliability check was carried out by a second rater, blind to any knowledge about the child and family, and rescored from audiotapes of 79 randomly selected interviews. The average intraclass correlations (ICC) were .98 for total number of ADHD symptoms, .94 for total impairment score of ADHD, .98 for total number of ODD symptoms, .85 for impairment score of ODD, .91 for total number of CD symptoms, and .99 for impairment score of CD. Covariates. Covariates included maternal and paternal educational level (socioeconomic status [SES]) obtained from the MoBa questionnaire at 17th gestational week. Information about maternal age at delivery and marital status were obtained from The Norwegian Medical Birth Registry (Magnus, Haug, Nystad, & Skjaerven, 2006). From the neuropsychological assessment performed by a trained psychologist, a short form of Stanford Binet 5th edition provided
information about general intellectual functioning (Roid, 2003). The Regional Ethics Committee and the Norwegian data inspectorate approved the study in 2007. Assessments were carried out according to ethical standards, and principles of the Helsinki declaration were followed. Parents returned a written consent prior to the clinical assessment.
Data Analysis Maternal, paternal, and offspring characteristics for all participants were compared according to presence of ADHD (yes/no) in the child. We examined statistical differences in background characteristic by independent-sample t tests for continuous measures and Pearson chi-square tests or Fisher’s exact tests for categorical measures. Covariates significantly associated with ADHD were adjusted for in the subsequent multiple regression analyses. We examined the odds ratios (ORs) of having co-occurring ODD and/or CD in ADHD by binary logistic regression and adjusted for possible confounding by gender, child’s age (months), IQ, maternal age (years), marital status (cohabiting/married or single parent), maternal education (years), and paternal education (years). We examined for possible confounding or interactions between independent variables (i.e., sex, ODD, and CD). The associations between the numbers of ODD and/or CD symptoms to the numbers of ADHD symptoms were examined by Pearson product–moment correlations and by hierarchical multiple linear regression analyses. There were no statistically significant non-linear relationships of ODD and CD symptoms, nor were there significant interactions between independent variables (i.e., sex, ODD, and CD symptoms) in predicting the number of ADHD symptoms. Comorbidity patterns of ODD and/or CD across different subtypes were compared by crosstabs and Mantel Haenzel’s ORs were reported. Furthermore, a one-way ANOVA between the groups was conducted to examine the relationship between comorbidity of ODD or CD and the severity of IA and HI symptoms. Participants were divided into four groups: ADHD only, ADHD plus ODD, ADHD plus CD, and ADHD plus ODD plus CD. Post hoc comparisons with Bonferroni correction were used. Finally, we examined sex distribution and sex differences in ADHD, subtypes of ADHD, and in co-occurring symptom clusters by independent-sample t tests for continuous measures and Pearson chi-square tests or Fisher’s exact tests for categorical measures. All tests were two-tailed.
Results Prevalence rate of children fulfilling the DSM-IV-TR diagnostic criteria for ADHD was 16% (n = 163). Table 1 shows
Downloaded from jad.sagepub.com at UNIV NEBRASKA LIBRARIES on October 7, 2015
5
Bendiksen et al. Table 1. Sample Characteristics of 1,048 Preschool Children in the ADHD Study.
ADHD (n = 163)
No ADHD (n = 885)
n (%)
n (%)
97 (59.5) Sex (boys) Age (months) IQ ADHD symptoms IA symptoms HI symptoms ODD symptoms CD symptoms Mothers educational level (years) Fathers educational level (years) Mothers age (years)
448 (50.6)
M (SD)
M (SD)
41.6 (1.3) 99.5 (9.0) 10.6 (2.8) 3.8 (2.4) 6.8 (1.3) 2.8 (1.9) 1.3 (1.3) 14.4 (2.5) 13.6 (2.9) 29.9 (4.4)
41.8 (1.3) 102.2 (9.2) 3.1 (2.7) 1.0 (1.3) 2.2 (1.9) 1.5 (1.4) 0.6 (0.9) 15.4 (2.3) 14.6 (2.7) 30.6 (4.2)
n (%)
Civil status (single parent)
9 (5.5)
n (%) 27 (3.1)
χ2/t test χ2(df)
p
4.4 (1)
.04
t(df)
p
−1.3 (1,044) −3.3 (1,033) 30.5 (1,046) 14.3 (181) 38.9 (322) 8.3 (198) 6.8 (186) −4.4 (209) −3.9 (205) −2.1 (1,045)
.19 .001
