Skip to content

Bipolar Disorder in Children and Adolescents: New Data to Inform Classification


Location: Washington, DC

Sponsored by:
National Institute of Mental Health (NIMH)
Division of Developmental Translational Research (DDTR)


Differences between research groups in the approaches used to diagnose bipolar disorder in children and adolescents have limited the comparability of findings across research studies and have impeded progress in understanding the course of illness, in delineating etiology, and in developing efficacious interventions. The purpose of this meeting was to take an important step forward in resolving these methodological discrepancies. The meeting had three specific goals:

  • Identify the differences among approaches currently used to diagnose bipolar disorder in youth and reach consensus on how to operationalize symptoms/criteria in research studies of Bipolar-I, -II, and -NOS
  • Identify approaches that can be used to reconcile differences and reach consensus as to the primary assessment tools to be used in research studies assessing bipolar youth
  • Define areas of classification or assessment that require further investigation

NIMH convened a working group of 24 invited experts to present and discuss data to address the meeting goals. This summary outlines the discussion and describes the guidelines developed by the working group.

Important Considerations

  • The working group's recommendations are for research purposes only and not for clinical/treatment decisions.
  • The recommendations provide definitions of criteria for mania and guidelines for defining an onset and a recovery from an episode of mania. While these guidelines are designed to elucidate a childhood phenotype of bipolar disorder, it is equally important to consider a broad range of alternative phenotypes to compare to these more strictly defined forms.
  • The recommendations are not meant to be definitive, authoritative, and fixed, but rather, expected to evolve with the further development of an empirical base.
  • Adopting these recommendations in research studies that apply labels of Bipolar-I, -II, and -NOS (BP-I, BP-II, BP-NOS) to children and adolescents is expected to enhance comparability of findings across research groups.
  • Investigators can collect additional data about relevant psychopathology to address remaining nosological questions. To maintain transparency and comparability across research studies, these data should be explicitly identified as exploratory in nature and not be used to alter the diagnosis of bipolar disorder as detailed below.

Goal 1: Consensus on Operationalizing Criteria and Symptoms

There was agreement that DSM-IV criteria be used to guide diagnoses of BP-I and -II in research studies of children and adolescents. Adherence to DSM-IV alone, however, is not sufficient to ensure consistency, as investigators have interpreted and applied criteria differently across sites. Current areas of divergence in the diagnosis of mania/hypomania include approaches for defining an episode; defining the threshold for episode duration; counting symptoms that overlap with other disorders, including irritability in the diagnostic process; and operationalizing grandiosity and elation in young children (not resolved at this meeting).

Defining an episode. The significant challenges to identifying episodes of mania in children—e.g., difficulty in establishing change from baseline to onset and offset of symptoms, complex cycling patterns, lack of clear periods of wellness/euthymia, varying assessment methods—have resulted in different approaches to defining episodes. Based on available data and expertise, the working group recommended the following:

  • An episode must meet full DSM-IV criteria: a sufficient number of Criterion B symptoms (3 symptoms; 4 if irritable mood only) must onset at the same time as the Criterion A change in mood; or, if the Criterion B symptoms are present at baseline, these symptoms must worsen concurrent with the change in mood. Thus, symptoms must be distinctly different from the child's baseline. In addition, symptoms must be different from age-appropriate behavior and not better accounted for by the presence of other psychiatric disorders.
  • An episode must meet DSM-IV duration criteria: 7 consecutive days for mania (less if child is hospitalized); 4 consecutive days for hypomania.
  • A new episode begins after full remission, but not after polarity switching.

Threshold of criteria duration. There was agreement that symptoms must be present for 4 total hours per day, more days than not.

Overlapping symptoms. A number of symptoms of mania/hypomania overlap with symptoms of other childhood disorders, and investigators differ on whether they "double count" symptoms across disorder categories. Many argue that seasoned clinicians can distinguish symptoms by category (e.g., ADHD hyperactivity versus mania symptoms, manic irritability versus depressive irritability); however, this contention has not been tested systematically. Additional challenges include possible ceiling effects (i.e., severity of symptoms may not worsen with onset of a manic episode, because the symptoms are already extreme) and differences in assessment approaches across diagnostic interviews (e.g., assess symptom in each diagnostic section or only some sections). To manage these obstacles, the working group made the following recommendation:

  • For a symptom to count toward the diagnosis of mania/hypomania, it must be anchored to the mood disturbance, i.e., either onset occurs at the same time as the mood disturbance or, if the symptom is present prior to the mood disturbance, intensifies at the same time that the mood disturbance begins.

Irritability. There is inconsistency in whether irritability has been counted toward Criterion A and whether chronic irritability is viewed as fulfilling Criterion A in children. The working group made the following recommendations:

  • Consistent with DSM-IV, irritable mood fulfills Criterion A.
  • Consistent with DSM-IV, irritable mood should be episodic (i.e., occurring with symptoms from Criterion B), or if chronic, worsening at the time of the episode.
  • Consistent with DSM-IV, 4 or more symptoms (during the worsening) from Criterion B must be present to meet full manic/hypomanic episode criteria if irritability is the only Criterion A mood symptom.

Sub-threshold BP Research Criteria. The working group also discussed BP-NOS, and considered cases that (a) meet Criterion A but fall short of full DSM-IV Criterion B symptoms or (b) meet full DSM-IV manic/hypomanic Criterion B symptoms but fall short of DSM-IV duration for hypomania. The working group discussed the latter subgroup of cases in more detail (as they appear more common in ongoing, phenomenology studies) and labeled this subgroup Sub-threshold BP. Suggestions for defining Sub-threshold BP are much more tentative than the guidelines for BP-I and -II described above:

  • Does meet DSM-IV Criterion B for manic or hypomanic episode (i.e., all the symptoms), except for duration criteria (< 4 consecutive days)
  • Four hours/day of manic symptoms to count as a day; require lifetime occurrence of > 20 days meeting DSM-IV criteria for mania or hypomania
  • There is a distinct change in functioning (but not necessarily functional impairment)
  • The episode is not substance- or medication-induced and symptoms are not better accounted for by other disorders
  • Does not meet DSM-IV criteria for cyclothymic disorder

The group agreed that future studies need to further explore Sub-threshold BP and other BP-NOS subgroups. Investigators should explicitly document how BP-NOS is defined in their studies.

Goal 2: Consensus on Primary Assessment Tool(s) for Clinical Research Studies

The three versions of the KSADS most widely used in clinical studies of bipolar disorder in children (KSADS-E, KSADS-PL, WASH-U-KSADS) vary in their approaches to assessing DSM-IV criteria. In addition, investigator decisions about how to apply DSM-IV rules influence the generation of diagnosis. For individual clinical research studies, the group acknowledged that available instruments could be adapted and blended to assess bipolar disorder as described above. For collaborative efforts, the group recommended that a combination of scales be used to assess mania; this would permit adequate assessment of individual symptoms and their onset, offset, and severity, and the overall onset and offset of the disorder to assess frequency of episodes. Many recommended that the KSADS-PL be used in combination with a modified form of K-MRS/K-DRS, or with the WASH-U-KSADS Affective Module, modified to require episodes. This type of approach has already been adopted by many recent, ongoing NIMH studies. The group also agreed to characterize bipolar disorder better, it was necessary to assess pervasive developmental disorders (PDD)—the KSADS-PL and WASH-U-KSADS do not include PDD symptoms—in addition to anxiety, ADHD, and other behavioral disorders.

Goal 3: Areas Requiring Further Investigation

The working group reached majority agreement in many important areas; however, there was unanimous agreement that more data are needed to further address these and other nosological issues. Future research directions include:

  • Conducting further empirical studies of nosology of BP-I using these definitions
  • Conducting further empirical studies of BP-II, NOS, cyclothymia, severe chronic irritability, and alternative phenotypes that best represent the range of behaviors presenting for treatment
  • Establishing a network to strengthen diagnostic consensus on operationalizing episodes and specific symptoms (e.g., grandiosity and elation) and on counting symptoms that overlap with other disorders
  • Developing common measures for collaborative studies
  • Mining existing data to evaluate rates of complete remission, incomplete remission, and recovery, and to determine phenotypes in children and adolescents that predict bipolar disorder in adulthood
  • Conducting research studies that quantify, qualify, and compare manic irritability to other forms of irritability
  • Developing better descriptions and developmentally-appropriate anchors for grandiosity, elation, expansive, and other manic symptoms in children
  • Examining mechanisms of comorbidity with anxiety, PDD, ADHD and other behavioral disorders
  • Evaluating distinguishing features of bipolar disorder, such as impairment in the regulation of positive affect, depression, behavioral activation, anger
  • Continuing to define clinical presentations based on shared underlying biological causes/pathways to further studies of pathophysiology
  • Characterizing the prodrome of bipolar disorder in children and adolescents and agree on diagnostic criteria for prodromal categories
  • Characterizing the pathology in children with first degree relatives with bipolar disorder
  • Studying the long-term course and outcome of pediatric bipolar spectrum disorders

For more information, please contact Shelli Avenevoli, Ph.D.,