- • Bipolar patients with a history of childhood attention-deficit hyperactivity disorder (ADHD) have a different course of illness regardless of whether they meet the ADHD criteria in adulthood or not.
Objective: The occurrence of comorbid attention-deficit hyperactivity disorder (ADHD) might have an impact of the course of the bipolar disorder.
Method: Patients with bipolar disorder (n = 159) underwent a comprehensive evaluation with respect to affective symptoms. Independent psychiatrists assessed childhood and current ADHD, and an interview with a parent was undertaken.
Results: The prevalence of adult ADHD was 16%. An additional 12% met the criteria for childhood ADHD without meeting criteria for adult ADHD. Both these groups had significantly earlier onset of their first affective episode, more frequent affective episodes (except manic episodes), and more interpersonal violence than the bipolar patients without a history of ADHD.
Conclusion: The fact that bipolar patients with a history of childhood ADHD have a different clinical outcome than the pure bipolar group, regardless of whether the ADHD symptoms remained in adulthood or not, suggests that it represent a distinct early-onset phenotype of bipolar disorder.
Bipolar disorder and attention-deficit hyperactivity disorder (ADHD) overlap in children and adolescents. Cross-sectional studies suggest that up to 85% of prepubertal children with bipolar disorder also meet the criteria for ADHD, and conversely, that up to 22% of children with ADHD also meet the criteria for bipolar disorder (1). In bipolar adults, the prevalence of comorbid ADHD has been less studied: The National Co-morbidity Survey found comorbid ADHD in 21% of patients with bipolar disorder (2). In the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD), the prevalence of comorbid ADHD was 9.5% (3).
The occurrence of comorbid ADHD might have an impact of the course of the bipolar illness. Sachs et al. (4) found that bipolar adults with a history of ADHD had earlier onset of disease than bipolar patients without ADHD. The STEP-BD corroborated this finding and also found that patients with comorbid ADHD had more manic and depressive episodes, more suicide attempts, more violence and substance abuse and greater likelihood to suffer from anxiety disorders (3).
The interpretation of these findings is, however, hampered by limitations with respect to the assessment of comorbid ADHD. Recently, Tamam et al. conducted a meticulous survey of comorbid ADHD in 159 bipolar adults, in which they rated self-reported childhood symptoms in addition to present symptoms of ADHD (5). They thereby identified a group of bipolar patients (10.7%) who reported a history of childhood ADHD but did not meet the criteria for adult ADHD. Interestingly, the clinical outcome in this group was comparable to the group with adult ADHD. However, although Tamam et al. used a validated rating scale to assess childhood ADHD symptoms, they relied solely on patient reports. As doubts have been cast on the accuracy of patients’ own recollections of attention-deficit symptoms, participant’s characteristics in adulthood might have affected the recall of childhood symptoms and confounded the results. One way to circumvent potential bias in terms of a participant’s own recollection of childhood symptoms (6, 7) is to supplement patient reports with parent interviews.
The question whether childhood ADHD has the same impact on bipolar disorder regardless of whether the ADHD symptoms remain in adulthood or not is critical, because it opens up the possibility that a mere history of childhood ADHD symptoms defines an etiologically distinct, early-onset bipolar sub-phenotype.
To study whether a history of childhood ADHD has an impact on clinical outcome in adult bipolar patients regardless of current ADHD.
We compared bipolar patients without a history of ADHD (pure bipolar) with: i) bipolar patients with a history of childhood ADHD who also fulfilled the DSM-IV criteria for current ADHD (adult ADHD) and ii) bipolar patients with a history of childhood ADHD who did not meet the DSM-IV criteria for current ADHD (childhood only ADHD). To increase the reliability of the childhood ADHD symptom rating, we conducted parent interviews to supplement patients’ self-reports.
The patients who participated in this report were recruited between October 2005 and April 2008 from a bipolar out-patient tertiary clinic at the Northern Stockholm psychiatric clinic, Stockholm, Sweden. This clinic serves northern Stockholm, an urban area with a total population of 316 400 persons over 18 years of age. The catchment area includes the entire spectrum of backgrounds, from wealthy upper class areas with a high proportion of native-born Swedes to ethnically diverse areas with high deprivation indices. All patients who present symptoms of mania, hypomania or other signs of bipolar disorder in this catchment area receive initial care at this clinic, and are referred to a tertiary care bipolar out-patient unit for work-up and further treatment. This means that virtually all new bipolar patients within the specified catchment area were referred for evaluation to this out-patient unit during the recruitment period. Both new consecutive out-patients who had been referred for treatment as well as continuing patients at the tertiary care bipolar out-patient unit were invited to participate, provided that they had been diagnosed with bipolar disorder.
The baseline clinical diagnostic instrument for bipolar disorder was the Affective Disorder Evaluation (ADE). The ADE was the diagnostic instrument used in the STEP-BD project (8). With the permission of the originator Gary S. Sachs, the ADE was translated and modified to suit Swedish conditions. The ADE starts with a social anamnesis, followed by the affective module of the Structured Clinical Interview for DSM-IV (9). The number of lifetime affective episodes and their characteristics are documented. Other modules assess alcohol and drug abuse, violent behaviour, childhood history, family history, treatment history, reproductive history and somatic illnesses. Interpersonal violence is defined as a violent act or serious physical threat to another person. Suicide attempt is defined as a deliberate and serious self-injury, including intoxication with medication, with the intent to die or seriously harm oneself. In addition to the ADE, the structured psychiatric interview Mini International Neuropsychiatry Interview (M.I.N.I) (10) was completed at baseline to screen for psychiatric diagnoses other than bipolar disorder. To screen for alcohol and substance abuse, patients completed two self-report questionnaires: the Alcohol Use Disorders Identification Test (AUDIT) (11) and the Drug Use Disorders Identification Test (DUDIT) (12). The ADE and M.I.N.I. interviews were conducted by board-certified psychiatrists working at the tertiary care bipolar out-patient unit, or by residents in psychiatry completing their psychiatric training at this unit. The assessments were based on all available sources of information, including patient records and interviews with next of kin where feasible. The patient data was then presented at a diagnostic case conference, where the final diagnostic decision was made by a consensus panel of experienced board-certified psychiatrists specialized in bipolar disorder. Patients who were at least 18 years of age and met the DSM-IV criteria for any bipolar disorder, i.e. type I, II, NOS, cyclothymia, or schizoaffective syndrome manic type, were asked to return to obtain detailed written and oral information about the study. Patients were excluded if they were unable to complete the standard clinical assessment or were incapable of providing informed consent. All enrolled patients consented orally and in writing to participate in the study.
After the bipolar diagnosis had been established, patients were referred to the assessment of childhood and adult ADHD symptoms; this was done at a point in time when the patient was not suffering from an acute affective episode and considered to be euthymic by the treating physician. The ADHD assessments were conducted by two independent, board-certified psychiatrists (E.R. and D.S.), working at a tertiary care unit specialized in the diagnosis and treatment of adult ADHD. Although these investigators were aware that all patients had been diagnosed with bipolar disorder, they were blind with respect to all other results from the baseline interview. Patients were informed that the purpose of the examination was to capture the typical functioning of the individual when he/she did not have affective symptoms. The ADHD clinical assessment required approximately 1.5 h. All rating scales and self-report questionnaires, except the parental interview, were completed during the same session. The interviewer then asked for permission to contact the patient’s parent by phone, or other next of kin if no parent was reachable. The telephone interviews were conducted according to a structured protocol, see below, and required approximately 30 min to complete.
The following instruments were used to assess various aspects of ADHD:
The 159 bipolar patients were divided into three groups based on the following definitions:
The computer software statistica release 7 was used in the statistical analyses. When comparing parametric data anova was used. Kruskal–Wallis anova was used when the variables were non-parametric. For cross-tabulation analyses, the Fisher exact test was used. Pearson correlation was used for all correlation analyses. Regression analyses were used to adjust for important clinical confounding factors. Multiple regression analysis was used to investigate if a history of childhood ADHD correlated to frequency of total episodes, depressive- and hypomania episodes, when adjusted for gender, bipolar subgroup and duration of affective illness. These outcome variables were used as dependent variables respectively, with gender, age, bipolar subgroup (BPI or BPII) and childhood ADHD as independent variables.
Logistic regression analysis was used when the outcome variables were dichotomous and episodes few. Logistic regression was used to investigate if a history of childhood ADHD increases the odds ratio of mixed episode, interpersonal violence or psychosis when adjusted for gender, bipolar subgroup and duration of affective illness. These outcome variables were used as dependent variables respectively, with gender, age, bipolar subgroup (BPI or BPII) and childhood ADHD as independent variables.
The kappa coefficient was calculated to rate the agreement between the self-assessment and parental assessment of childhood ADHD. A two-sided P-value of <0.05 was considered statistically significant.
The cohort consisted of 159 bipolar patients (60 males and 99 females), whereof 80 were diagnosed as bipolar I, 64 as bipolar II and 15 as bipolar disorder NOS. The mean (SD) age was 39.1 (12.8) years. The prevalence of pure bipolar disorder without a history of ADHD was 71.7% (n = 114; 95% CI: 64.6–78.8). There were no significant differences with respect to gender distribution or age between the adult ADHD, childhood only ADHD and pure bipolar groups (Table 1). Present alcohol consumption did not differ between the groups as measured by AUDIT (mean points ± SD), adult ADHD 2.7 ± 1.1; childhood only ADHD 2.2 ± 1.3; pure bipolar 3.5 ± 0.5, H = 1.3, P = 0.59.
|Pure bipolar, n = 114||Adult ADHD, n = 26||Childhood only ADHD, n = 19||Statistics|
|P-value||Comparison between groups|
|Age at first psychiatric symptom||21.9||11.1||15.4||11.8||12.7||5.7||8.6||<0.001||1 > 2, 1 > 3|
|Age at first affective episode||23.4||10.3||18.2||8.7||16.6||6.2||6.8||0.014||1 > 2, 1 > 3|
|Psychotic features||63||55.3||5||19.2||8||42.1||12.0||0.0027||1 > 2|
|Suicide attempt||35||30.1||14||53.9||6||31.6||4.8||0.093||1 < 2|
|Interpersonal violence||20||17.5||11||42.3||8||42.1||10.0||0.0068||1 < 2, 1 < 3|
|Frequency of episodes|
|Hypomania||3.6||6.4||15.4||20.8||8.8||12.2||19.9||<0.001||1 < 2|
|Mixed||0.81||3.4||4.2||9.1||5.6||9.2||19.7||<0.001||1 < 3|
|Total||13.4||14.3||33.4||33.7||24.6||25.6||12.2||0.0022||1 < 2|
The adult ADHD group comprised those who fulfilled the DSM-IV criteria for any type of ADHD in adulthood along with a history of childhood ADHD. The prevalence of any type of adult ADHD in this bipolar cohort was 16.4% (n = 26; 95% CI: 10.5–22.2), while 6.9% (n = 11; 95% CI: 2.9–10.9) met the criteria for the combined subtype (both inattention and hyperactivity-impulsivity) of ADHD. The adult ADHD group (any type) was significantly younger at first psychiatric symptom and first affective episode, compared to the pure bipolar group (Table 1). They also had significantly more hypomanic, and total affective episodes, as well as more suicide-attempts and interpersonal violent incidents (Table 1).
The childhood only ADHD group comprised those with a history of childhood ADHD who did not fulfil the DSM-IV criteria for any type of current adult ADHD. The prevalence of this group was 12.0% (n = 19; 95% CI: 6.9–17.0). In accordance with the adult ADHD group, this group was significantly younger at their first psychiatric symptoms, and was also significantly younger at their first affective episode compared to the pure bipolar group (Table 1). The childhood only ADHD group also had significantly more mixed episodes as well as significantly more interpersonal violent incidents, compared to the pure bipolar group (Table 1).
The results of the ADHD assessment scales are displayed in Table 2. The two assessments of childhood ADHD used in this study, WURS-25 and A-TAC (total ADHD score), were moderately correlated (r = 0.55, P < 0.001). The kappa value was 0.76, indicating good agreement between these two scales.
|Pure bipolar, n = 114||Adult ADHD, n = 26||Childhood only ADHD, n = 19||Statistics||P-value||Comparison between groups|
|Current ADHD measures|
|ASRS, total score||23.0||1.0–63.0||42.5||15.0–64.0||31.0||16.0–49.0||54.9||<0.001||1 < 2, 1 < 3, 2 > 3|
|Organizing/activating||8.0||0.0–25.0||18.0||9.0–26.0||13.0||2.0–25.0||40.4||<0.001||1 < 2, 1 < 3, 2 > 3|
|Sustaining attention/concentration||6.0||0.0–24.0||17.5||9.0–24.0||11.0||2.0–22.0||45.6||<0.001||1 < 2, 1 < 3, 2 > 3|
|Energy/effort||4.0||0.0–18.0||14.9||5.0–22.0||10.0||3.0–17.0||45.8||<0.001||1 < 2, 1 < 3, 2 > 3|
|Managing affective inferences||5.0||0.0–16.0||13.0||6.0–20.0||9.0||2.0–19.0||44.5||<0.001||1 < 2, 1 < 3, 2 > 3|
|Working-memory/access recall||3.0||0.0–15.0||8.5||0.0–17.0||5.0||0.0–13.0||32.6||<0.001||1 < 2, 1 < 3, 2 > 3|
|Total score||27.5||1.0–82.0||72.0||36.0–9.0||48.0||15.0–79.0||55.0||<0.001||1 < 2, 1 < 3, 2 > 3|
|Childhood ADHD measures|
|WURS-25||15.0||0.0–42.0||53.0||21.0–74.0||50.0||25.0–64.0||74.4||<0.001||1 < 2, 1 < 3|
|Attention||0.0||0.0–5.0||4.75||0.0–11.5||3.5||0.0–8.0||42.9||<0.001||1 < 2, 1 < 3|
|Impulsivity/activity||0.0||0.0–7.5||3.75||0.0–9.5||2.25||1.0–11.0||42.0||<0.001||1 < 2, 1 < 3|
|Planning/organizing||0.0||0.0–2.0||2.5||0.0–6.0||1.5||0.0–6.5||46.7||<0.001||1 < 2, 1 < 3|
The self-assessment of childhood ADHD (WURS-25) correlated significantly (P < 0.001) with both scales measuring current ADHD related problems: ASRS (r = 0.69) and BROWN ADD (r = 0.68). Likewise, the parental assessment of childhood ADHD (A-TAC) correlated significantly to both ASRS (r = 0.63) and BROWN ADD (r = 0.51).
Given that adult ADHD and childhood only ADHD had similar impact on the clinical outcome, we merged these two groups into one: bipolar patients with a history of childhood ADHD regardless of current ADHD. These patients were then compared with the pure bipolar group (Table 3).
|Childhood ADHD, n = 45||Pure bipolar, n = 114|
|Occurrence of any episode||n||%||n||%||OR (95% CI)||Adjusted P-value*|
|Mania episode||19||42.2||61||53.5||1.54 (0.76–3.12)||0.22†|
|Mixed episode||16||35.6||12||10.6||5.2 (2.0–13.6)||<0.001|
|Interpersonal violence||19||42.2||20||17.5||4.6 (1.9–11.2)||<0.001|
|Psychotic feature||13||28.9||63||55.3||0.25 (0.09–0.70)||0.008|
|Number of episodes||Mean||SD||Mean||SD||B (SE)||Adjusted P-value*|
The total prevalence of a history of childhood ADHD was 28.3% (n = 45; 95% CI: 21.2–35.4), but it was significantly more prevalent in the bipolar II group than in the bipolar I group (37.5% vs. 23.8%; P = 0.048). These patients were younger when experiencing their first psychiatric symptoms (14.2 vs. 21.9 year; P < 0.001) and when experiencing their first affective episode (16.7 vs. 22.7 year; P < 0.001) than the pure bipolar group. They had significantly more hypomanic, depressive mixed and total affective episodes, and they more often had a history of interpersonal violence. Suicide attempts was numerically, but not significantly, more common (44.4% vs. 30.7%; P = 0.08) in those with a history of childhood ADHD. They were, however, less likely to have had psychotic features than the pure bipolar group. The groups did not differ with respect to the occurrence of any manic episode or the number of manic episodes.
We performed multiple regression analyses to adjust for the confounding factors gender, duration of affective illness and bipolar subgroup. These analyses confirmed that bipolar patients with a history of childhood ADHD were at significantly higher risk for more frequent episodes of hypomanic, mixed, depressive, and total episodes, along with more frequent interpersonal violence and less psychotic features (Table 3).
This is one of the most rigorous assessments to date of comorbid ADHD, including childhood ADHD, in adult bipolar patients. Not only were patients assessed by MDs specially trained in the assessment of bipolar disorders, but they were also independently assessed by psychiatrists at a unit specialized in the diagnosis and treatment of adult ADHD. Finally, an interview with a parent was conducted to obtain objective information about a history of childhood ADHD.
The prevalence of a history of childhood ADHD in combination with present ADHD was 16.4%. To meaningfully compare the prevalence figures of comorbid ADHD across studies, it is important to consider the definitions used to diagnose ADHD. The prevalence figures in this study can be compared with the study of Tamam et al. that used similar definitions and estimated the prevalence to 16.3% (18). The total prevalence of those with a history of childhood ADHD, regardless of current ADHD, was 28.3% in the present study and 27.0% in the Tamam study. The STEP-BD study found comorbid ADHD in only 9.5% in their cohort (3). However, according to their definition of ADHD, this figure should be compared with the prevalence of the combined subtype of ADHD in the present study which was 6.9%. Likewise, the National Co-morbidity Survey (2) reported ADHD in 21.2% of their bipolar patients based on a less strict definition of childhood ADHD that would yield 22.4% ADHD in our sample. Hence, the prevalence of comorbid ADHD is consistent across studies provided that the same definitions are employed.
Both A-TAC and WURS are easy-to-use tools that might be used to capture a history of childhood ADHD. We found that combining self-assessment (WURS-25) with a parental interview of childhood ADHD problems (A-TAC) captured more bipolar patients with childhood ADHD than self-report only. However, conducting the self-assessment (WURS-25) is probably sufficient for clinical purposes, given that this instrument correlated to the parental interview.
We found that bipolar patients with adult ADHD had higher frequency of affective episodes (except manic episodes) than pure bipolar patients. This accords previous surveys (3, 18). More importantly, however, the group of bipolar patients with childhood ADHD who did not fulfil criteria for ADHD in adulthood had a similar clinical course, which differed from those with pure bipolar disorder. Bipolar disorder with comorbid ADHD has previously been suggested to represent a distinct bipolar phenotype (1, 19, 20), but our findings suggest that a mere history of childhood ADHD – irrespective of current ADHD – might represent a specific sub-type of bipolar disorder.
A mixture of ADHD and bipolar symptoms might, however, also represent the co-occurrence of two distinct illnesses as suggested by a recent MRI study (21). It is conceivable that the persistent dysregulation of activity level and affects in-patients with ADHD might trigger more frequent bipolar episodes. This study cannot elucidate these speculations; further studies of biological markers in bipolar subjects with and without a childhood history of ADHD is warranted to shed light on this issue.
There are some facets of this study to consider when interpreting the results. First, recall bias is a general hurdle in retrospective ADHD diagnosing. Even though we tried to minimize this effect by interviewing parents, we cannot rule out that some recall bias might have occurred. Moreover, parental interviewing was only possible in 2/3 of the subjects, most often because of lack of consent from patients or because of the lack of relatives with knowledge about the patient’s childhood. Second, it might be argued that the symptoms of some patients might be better accounted for by an ADHD diagnosis only, rather than a combination of ADHD and bipolar disorder. This risk has, however, been minimized in this study by an extensive initial evaluation of bipolar disorder certifying that diagnostic criteria were fulfilled. Third, some patients missed scheduled appointments for ADHD assessment. In fact, 30 patients were scheduled more than three times. In our experience, missing appointments might be a sign of ADHD per se. The missed appointments might therefore contribute to an underestimation of the true prevalence of comorbid ADHD. There were also subjects who did not want to participate in the ADHD evaluation and a few who could not be scheduled because they did not recover from affective symptoms. Fourth, although the physician deemed the patients to be euthymic, and rating scales of mania and depression were used to ensure euthymia in conjunction with other assessment, no structured rating of depression and mania was conducted in immediate conjunction with the ADHD assessment. However, as patients were explicitly asked to describe their normal functioning when they were free from affective symptoms to capture true ADHD symptoms, it is unlikely that acute affective symptoms have confounded the results.
In conclusion, we found that childhood ADHD is common in bipolar patients and that a mere history of childhood ADHD irrespective of current ADHD is an important factor for the clinical course of bipolar disorder. A review of childhood ADHD symptoms in adult bipolar patients is warranted both clinically and in pathophysiological studies.
The authors would like to thank study coordinator Martina Wennberg for skilful assistance, and Dr Caroline Nilsson for the assessment of bipolar patients. Financial support was provided through the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and the Karolinska Institutet; and through grants from the Swedish Medical Research Council (K2008-62x-14647-06-3), the S:t Göran foundation, the Söderström-Königska Foundation, the Thuring Foundation, the Swedish Psychiatry foundation, the Swedish Society for Medical Research, and the Karolinska Institutet.