Article

Psychiatric Comorbidity in Adult Attention Deficit Hyperactivity Disorder: Findings From Multiplex Families James J. McGough, M.D. Susan L. Smalley, Ph.D. James T. McCracken, M.D. May Yang, M.S. Melissa Del’Homme, Ph.D. Deborah E. Lynn, M.D. Sandra Loo, Ph.D.

Objective: Patterns of psychiatric comorbidity were assessed in adults with and without attention deficit hyperactivity disorder (ADHD) identified through a genetic study of families containing multiple children with ADHD. Method: Lifetime ADHD and comorbid psychopathology were assessed in 435 parents of children with ADHD. Rates and mean ages at onset of comorbid psychopathology were compared in parents with lifetime ADHD, parents with persistent ADHD, and those without ADHD. Age-adjusted rates of comorbidity were compared with Kaplan-Meier survival curves. Logistic regression was used to assess additional risk factors for conditions more frequent in ADHD subjects. Results: The parents with ADHD were significantly more likely to be unskilled workers and less likely to have a college degree. ADHD subjects had more lifetime psychopathology; 87% had at least one

and 56% had at least two other psychiatric disorders, compared with 64% and 27%, respectively, in non-ADHD subjects. ADHD was associated with greater disruptive behavior, substance use, and mood and anxiety disorders and with earlier onset of major depression, dysthymia, oppositional defiant disorder, and conduct disorder. Group differences based on Kaplan-Meier age-corrected risks were consistent with those for raw frequency distributions. Male sex added risk for disruptive behavior disorders. Female sex and oppositional defiant disorder contributed to risk for depression and anxiety. ADHD was not a significant risk factor for substance use disorders when male sex, disruptive behavior disorders, and socioeconomic status were controlled. Conclusions: Adult ADHD is associated with significant lifetime psychiatric comorbidity that is not explained by clinical referral bias. (Am J Psychiatry 2005; 162:1621–1627)

A

ttention deficit hyperactivity disorder (ADHD) is a highly heritable neurodevelopmental syndrome with significant lifetime risk for functional impairment (1). Coexisting psychopathology is common (2), and the potential clinical importance of comorbidity has been recognized in children (3). Our understanding of psychiatric comorbidity in adult ADHD is based on a limited number of reports. Changes in ADHD nosology, increased awareness of ADHD persistence, and differing diagnostic approaches to adults further confound our appreciation of the relationship between adult ADHD and other disorders. In this current report we examine ADHD and comorbid psychopathology in a group of adults ascertained through pairs of siblings with ADHD. Several longitudinal studies initially informed our understanding of lifetime psychopathology in hyperactive children. From an original group of 101 hyperkinetic boys compared with nonhyperactive comparison subjects, investigators described higher rates of conduct disorder, antisocial personality disorder, and nonalcoholic substance use disorders in the probands at early adult follow-up than in the comparison subjects (4). Marijuana and cocaine were the most frequently used recreational drugs, and initiation of substance abuse was generally subsequent to conduct disorder. Mood and anxiety disorders were rare; Am J Psychiatry 162:9, September 2005

however, the subjects had not fully passed through the risk periods for these conditions. Additionally, subjects were excluded if the primary reason for referral was aggression. A second cohort of hyperkinetic children showed significantly more antisocial personality disorder and somewhat higher rates of alcohol and marijuana use in early adulthood than did comparison subjects (5). These studies might have underestimated some comorbidities because of the lack of follow-up for some subjects. However, higher rates of antisocial behavior and substance use disorders emerge as consistent findings in virtually all studies of comorbidity and adult ADHD. A more complicated picture has been described in patients diagnosed with ADHD as adults. Shekim and colleagues (6) assessed 51 clinically referred ADHD adults and found high rates of lifetime comorbidity, including 51% for generalized anxiety disorder, 34% for alcohol abuse or dependence, 34% for other drug abuse or dependence, 25% for dysthymia, 18% for separation anxiety disorder, 13% for obsessive-compulsive disorder (OCD), and 10% for major depressive disorder. Downey and colleagues (7) described psychological profiles for 78 clinically referred ADHD adults and reported that 37% evidenced current depression, 33% had alcohol abuse or dependence, 22% had http://ajp.psychiatryonline.org

1621

COMORBIDITY IN ADULT ADHD TABLE 1. Characteristics of Parents in Families Containing Multiple Siblings With ADHD Parental Characteristic Age (years) Fathers (N=208) Mothers (N=227) Total (N=435) Race White Nonwhite Education College or above Below college degree Marital status Intact marriage Nonintact marriage

Mean

SD

Range

44.7 41.9 43.2

5.9 5.1 5.7

27–66 26–53 26–66

N

%

369 66

85 15

228 207

52 48

357 78

82 18

other drug abuse or dependence, and 13% had antisocial personality. These studies were uncontrolled and based on groups of treatment-seeking patients. Biederman and colleagues (8) studied 84 clinic-referred ADHD adults, 36 nonreferred ADHD adults, and 207 adults without ADHD. The ADHD groups demonstrated higher rates of oppositional defiant disorder, conduct disorder, depression, substance abuse/dependence, and anxiety. Subsequent reports described a higher risk for psychoactive substance use disorders, particularly among men and subjects in lower socioeconomic classes (9). The ADHD individuals had earlier onsets of substance use disorders than did the comparison subjects (10, 11). Murphy and Barkley (12), in a study of 172 clinically referred ADHD adults and 30 comparison subjects, noted higher rates of alcohol abuse or dependence (35% versus 10%), drug abuse or dependence (14% versus 3%), oppositional defiant disorder (30% versus 7%), and conduct disorder (17% versus 0%) in the patients with ADHD. The ADHD group also exhibited high rates of comorbid anxiety disorders (32%), major depressive disorder (18%), and dysthymia (32%), although these were not significantly higher than the rates for the comparison subjects. Several studies have examined differences in psychiatric comorbidity across ADHD subtypes. One study, using proxy DSM-IV subtypes in 149 adults diagnosed according to DSM-III-R, found higher rates of oppositional defiant disorder, bipolar disorder, and substance use disorders in patients with the combined subtype than in those with other subtypes and higher rates of oppositional defiant disorder, OCD, and posttraumatic stress disorder (PTSD) in the hyperactive-impulsive subtype than in the inattentive subtype (13). A second study (14) found that subjects with the combined or hyperactive-impulsive subtype had a higher rate of oppositional defiant disorder than did subjects with the inattentive subtype, with no subtype difference in conduct disorder, depression, or anxiety. While existing evidence suggests that there is a significant risk of lifetime psychiatric comorbidity in adult ADHD, the design features of previously reported studies limit the generalizability of the findings. Studies based on DSM-III-R criteria might underrecognize comorbidity with the predominately inattentive subtype. Older studies

1622

http://ajp.psychiatryonline.org

of hyperactive children emphasized male subjects and might not represent clinical phenomenology in female subjects. In comparisons of comorbidity across ADHD subtypes, classification has been based on adult presentations, which do not reflect changes in symptom pattern over time. Finally, most published studies have been based on clinical study groups, which might bias the reported rates of psychopathology. The UCLA ADHD Genetic Study examined adult ADHD and comorbid psychopathology in a large, nonclinical study group. Families were recruited on the basis of having two or more affected siblings with ADHD. Parents of these affected sibling pairs underwent comprehensive assessment to identify lifetime psychopathology and ADHD versus non-ADHD status. We explored several specific hypotheses, namely, 1) that findings in ADHD adults identified through multiplex families would be similar to those for clinical groups described in the literature, specifically, that subjects with ADHD would have higher rates of psychiatric comorbidity; 2) that ADHD subjects would evidence earlier ages at onset for comorbid conditions; 3) that differences in comorbidity would occur across ADHD subtypes; and 4) that other, non-ADHD factors would contribute additional risk for psychopathology.

Method Study Group The subjects in this study were 435 parents of ADHD children ascertained through 230 families recruited by sampling of affected sibling pairs. The procedures for family recruitment, screening, and child assessment are detailed in a previous publication (15). All families spoke English and had biological parents available to participate. Families were included if at least one child met the full DSM-IV ADHD criteria and another child had a diagnosis of definite or probable ADHD. Probable ADHD was defined as a subject falling one symptom or criterion short of a full diagnosis (including behavioral symptoms, onset by age 7, duration of 6 months, or presence in two settings) but having evidence of impairment. Subject characteristics are summarized in Table 1. The mean socioeconomic status rank as defined by Hollingshead (16) was 2.3 (SD=0.9) (Hollingshead range=I–V).

Instruments and Procedures After receiving a full oral explanation of all study requirements and procedures, the parents and children provided written statements of informed consent/assent approved by the UCLA institutional review board. Each adult parent completed the Wender Utah Rating Scale as a self-report of childhood ADHD symptoms (17) and an ADHD Rating Scale-IV (18) as a self-report of current ADHD symptoms, as well as an ADHD Rating Scale-IV for current ADHD symptoms in the spouse. Clinical psychologists or highly trained interviewers with extensive experience and reliability training conducted the adult interviews. We determined lifetime history of psychopathology with the Schedule for Affective Disorders and Schizophrenia—Lifetime Version Modified for the Study of Anxiety Disorders (SADS-LA), a highly reliable semistructured diagnostic instrument for DSM-IV disorders (19). The SADS-LA was supplemented with the behavioral disorders section of the Schedule for Affective Disorders and Schizophrenia for SchoolAge Children—Present and Lifetime Version (K-SADS-PL) (20) to elucidate lifetime and persistent diagnoses of ADHD, as previously Am J Psychiatry 162:9, September 2005

McGOUGH, SMALLEY, McCRACKEN, ET AL. TABLE 2. Classification of ADHD Status for Parents in Families Containing Multiple Siblings With ADHD Definite Status Childhood/ lifetime ADHD

Persistent ADHD

Probable

DSM-IV Criteria Full retrospective DSM-IV childhood criteria

Other Characteristics DSM-IV Criteria Childhood impairment in two Lacking one retrospective or more settings, six or childhood symptom more childhood inattentive or criterion but with or hyperactive/impulsive evidence of symptoms before age 12, impairment some evidence of onset at or before age 7 — Full current DSM-IV Definite lifetime ADHD criteria with evidence of impairment

described (8). The K-SADS-PL was also used to elucidate retrospective diagnoses of childhood oppositional defiant disorder and conduct disorder. Our criteria for designation of lifetime and persistent ADHD are summarized in Table 2. Subjects who did not meet the criteria for lifetime ADHD constituted the non-ADHD group. Classification of the lifetime ADHD subtype was based on childhood symptoms between the ages of 5 and 12 years. Classification of persistent ADHD was based on current adult symptoms. “Best-estimate” diagnoses were assigned after review of all available clinical information, including semistructured interviews, behavioral ratings, and evidence of impairment, with senior clinicians (J.J.M., J.T.M., D.E.L.). All interviews were videotaped, and a subset was rerated independently by senior clinicians to maintain ongoing reliability. The mean weighted kappa based on 24 rerated adult tapes was 0.88 (SD=0.10), with weighted kappa values of 0.91, 0.86, and 0.83 for ADHD, oppositional defiant disorder, and conduct disorder, respectively (15). The mean kappa value computed for the agreement between the initial rater and the best-estimate diagnoses for 10 major psychiatric diagnoses occurring in more than 5% of all subjects in the current study group was 0.95 (SD=0.03).

Statistical Analyses Descriptive statistics were summarized. ADHD status was based on the criteria for either a probable or definite diagnosis. Diagnosis of other psychiatric conditions required DSM-IV criteria. We used t tests or chi-square analyses to determine differences in demographic information between groups. Statistically significant differences in demographic variables were assessed at an alpha level of