|Home | About | Journals | Submit | Contact Us | Français|
Bipolar Disorder–Not Otherwise Specified (BP-NOS) and Severe Mood Dysregulation (SMD) are severe mood disorders that were both defined to address questions about the diagnosis of bipolar disorder (BD) in youth. SMD and BP-NOS are distinct phenotypes that differ in clinical presentation and longitudinal course. The purpose of this review is to inform clinicians about the clinical features of the two phenotypes and about the research literature distinguishing them.
We review the literature on SMD as studied in the National Institute of Mental Health Intramural Research Program and on BP-NOS in youth. For BP-NOS, we focus on the phenotype defined in the Course of Bipolar Youth (COBY) study, since this has received the most study.
SMD is characterized by impairing, chronic irritability without distinct manic episodes. Most commonly, BP-NOS is characterized by manic, mixed or hypomanic episodes that are too short to meet DSM-IV-TR duration criterion. Research provides strong, albeit suggestive, evidence that SMD is not a form of BD; the most convincing evidence are longitudinal data indicating that youth with SMD are not at high risk to develop BD as they age. The BP-NOS phenotype appears to be on a diagnostic continuum with BD type I and type II. BP-NOS and BD- type I subjects have similar symptom and family history profiles, and youth with BP-NOS are at high risk to develop BD as they age. Currently, little research guides treatment for either phenotype.
Pressing research needs include identifying effective treatments for these phenotypes, ascertaining biomarkers that predict conversion from BP-NOS to BD, elucidating associations between SMD and other disorders, and defining the neural circuitry mediating each condition.
Bipolar Disorder–Not Otherwise Specified (BP-NOS) and severe mood dysregulation (SMD) are descriptive, heterogeneous phenotypic categories. Given the high prevalence and marked severity of these clinical presentations, clinicians are likely to encounter children with BP-NOS or SMD in their practices. However, despite cross-sectional and longitudinal differences between SMD and BP-NOS,1, 2 clinicians may believe that these are interchangeable terms for the same phenotype, or may be unclear as to how to differentiate them. The uncertainty may also stem from a view that BP-NOS, SMD, and bipolar disorder (BD) are part of one spectrum of conditions. The purpose of this review is to spotlight the salient clinical features of these conditions and their differences, since the diagnosis determines both the treatment plan and prognosis that will be communicated to patient and parents. To offer guidance on these conditions and to promote a shared vocabulary for the field, this paper a) reviews clinical criteria for BD, BP-NOS, and SMD; b) identifies clinical features that distinguish the three conditions; and c) reviews literature supporting the hypothesis that BP-NOS and SMD are distinct phenotypes. We end by suggesting topics for future research.
An understanding of the distinction between SMD and BP-NOS begins with understanding the definition of a manic episode in DSM-IV-TR. During a manic episode the child’s mood (elation, irritability, or both) changes to a level that is an obvious departure from what is typical for that child. When the change in mood has a sufficient duration; when its onset is concurrent with the start, or worsening, of other manic symptoms (e.g., decreased need for sleep, increased goal directed activity, distractibility); and when the symptoms impair functioning and are not due to other disorders, the diagnosis of a manic episode is clear. Thus, when these symptoms are episodic, the diagnosis of BD can be straightforward.
Often SMD and BP-NOS are considered when the clinical picture is not so straightforward. Both are more common than pediatric BD.2–6 Indeed, clinical studies report that youth who meet criteria for BP-NOS7 or SMD1 are at least as severely impaired as those with BD.
The DSM-IV definition of BP-NOS is very broad and non-specific i.e., “The Bipolar Disorder Not Otherwise Specified category includes disorders with bipolar features that do not meet criteria for any specific Bipolar Disorder.8 The text offers examples such as, “very rapid alternation (over days) between manic symptoms and depressive symptoms that do not meet minimal duration criteria for a Manic Episode or a Major Depressive Episode.8(p366) The most explicit definition of BP-NOS derives from the Course and Outcome of Bipolar Illness in Youth study (COBY).2 In this paper we have relied on COBY BP-NOS because it is the focus of a large, prospective, multisite study that compares BP-NOS to BD-I over an extended follow-up period. Although there are other BP-NOS definitions (see Table 1, Axelson et al., 2011),2 COBY BP-NOS is the most thoroughly investigated BP-NOS phenotype at this time. The definition of BP-NOS in COBY emphasizes that it is an episodic mood disorder, like BD. In its most common manifestation, children with COBY BP-NOS (hereafter referred to as BP-NOS) have a history of episodes that meet DSM-IV symptom criteria for mania or hypomania but are of shorter duration than the four days required by DSM-IV for a hypomanic episode7. SMD, which is not in the DSM-IV, is a prepubertal-onset phenotype characterized by chronic, severe, non-episodic irritability and hyperarousal symptoms.9 Of note, the designs of the investigations underlying SMD and BP-NOS are different and thus draw on different kinds of data. SMD has been studied at a single site but through many different studies, while the COBY study is a single study conducted in several sites.
No study has compared youth with SMD and those with BP-NOS directly. Consequently, for researchers and clinicians alike, the distinction between SMD and BP-NOS (as well as between each of these and BD) may pose diagnostic and therapeutic challenges. Indeed, the current American Academy of Child and Adolescent Psychiatry (AACAP) guidelines10, 11 define BP-NOS so broadly that it encompasses both BP-NOS and SMD. This is regrettable because emerging research, reviewed below, suggests that BP-NOS and SMD differ in clinical features, longitudinal course and family history.
At the time of writing, DSM-5 is being prepared. Draft documents propose a new diagnosis, Disruptive Mood Dysregulation Disorder (DMDD). In addition, there is considerable discussion about how hypomanic episodes shorter than 4 days in duration should be diagnosed in DSM-5. The proposed DSM-5 criteria for these syndromes are similar, but not identical to, the SMD and BP-NOS phenotypes discussed here. In particular, the proposal for DMDD, which arose largely from work on SMD, has generated considerable interest. Differences between the criteria for DMDD and those for SMD are detailed below, but it is important to note that the core feature of the two disorders i.e., non-episodic, severe, and impairing irritability, is the same. If DMDD is adopted in DSM-5, it will supplant SMD in both clinical and research settings. However, even if the corresponding DSM-5 diagnoses differ somewhat from the SMD and BP-NOS phenotypes discussed here, it is nonetheless important for clinicians to understand how to differentiate these two families of disorders i.e. those that, like BP-NOS/short-duration hypomania, are characterized by abbreviated but distinct episodes of manic symptoms vs. those that, like SMD/DMDD, are characterized by severe, chronic irritability. Drawing as it does on the largest body of work currently available on this topic, a review of the existing literature regarding BP-NOS and SMD is therefore an important avenue for educating clinicians about this differentiation, as well as about the possible treatment implications of this distinction and important future avenues of research.
BD is the reference point for explaining differences among it, SMD, and BP-NOS. DSM-IV criteria for BD (hereafter referring to BD type I or type II) are the same for adults and children, although developmental considerations are important when evaluating symptoms in children. A manic episode is defined by both a) a distinct period of elevated, expansive, or irritable mood, and b) associated features that have their onset, or significantly worsen during the period of abnormal mood.12
The critical defining feature of an episode is a predominant mood that is distinct from the child’s usual mood state and differs markedly from what is expected for the child’s developmental stage. During the episode, the child’s mood may vary but, most of the day, most days, it should differ clearly from that child’s baseline. By DSM-IV convention, a manic episode persists for at least 7 days and/or is severely impairing, leads to hospitalization, or includes psychosis; a hypomanic episode must last at least four days and be noticeable to others.8
Having first established a period of abnormal mood, the secondary defining feature of a manic episode is the presence of associated symptoms (“B-criteria”) that include inflated self-esteem or grandiosity, decreased need for sleep, more talkative than usual, pressured speech, flight of ideas, racing thoughts, distractibility, increased goal-directed activity, psychomotor agitation, and/or excessive involvement in pleasurable activities. It is crucial that B-criteria symptoms emerge or worsen markedly during a period of elevated or irritable mood. When B-criteria symptoms have their onset at about the same time as the change in mood, it is obvious that they should “count toward” the diagnosis. However, when a B-criterion symptom is also part of a concurrent disorder, it should be considered to be a symptom of mania only if it worsens significantly during the mood episode. In other words, chronic, pre-existing symptoms that do not worsen should not “count” toward a manic episode. For example, distractibility in a child with attention-deficit/hyperactivity disorder (ADHD) should be considered a symptom of a manic episode only if the distractibility is markedly more severe than usual during the time that he/she experiences irritable or euphoric mood that is also unusual for him/her.
The criteria for mania in DSM-IV8 do not address developmental considerations. The text comments only that substance abuse, truancy, antisocial behavior and psychotic symptoms are associated features in adolescents. Furthermore, the DSM-IV language, if taken concretely, frames some manic symptoms in ways that are inappropriate for children. For example, in pre-pubertal and early adolescent children, “excessive involvement in pleasurable activities with a high potential for harm” is unlikely to consist of “engaging in unrestrained buying sprees, sexual indiscretions, or foolish business investments.” However, children might attempt to buy expensive items on-line, become engrossed in sexually explicit websites, or attempt unrealistic business ventures. Also, to avoid “pathologizing” ordinary responses to exciting events, symptoms must be viewed in light of what is typical both developmentally and for that child. For example, elevated mood that accompanies or anticipates thrilling activities (e.g. birthdays, trips to theme parks) is rarely pathological.
Grandiosity, one of the prime symptoms of BD (including BP-NOS) but not of SMD, must be viewed in light of developmental and individual context. Children commonly may express unrealistic aspirations and elevated self-esteem that are consonant with their developmental level. For example, pre-teens and adolescents often overstate their own skills or abilities and may disparage adults in positions of authority. Some children habitually make such statements. On the other hand, such behavior takes on a different diagnostic meaning when it occurs in episodes of psychotic thinking and/or a qualitative change from that child’s usual statements and self-image. It is also useful to gauge the significance of such statements by how closely associated they are with out-of-character behaviors.
In summary, the key question facing the practitioner is whether, during the putative mood episode, the child’s mood and behavior were “abnormally” and “persistently” different across settings. Deciding whether a child has experienced a manic or hypomanic episode requires the clinician to use retrospective and prospective observation to understand the child’s usual functioning and then decide whether the child’s mood and behavior were qualitatively different during the suspected episode. This is particularly important for children with developmental or disruptive disorders. A practitioner who compares a child’s thoughts or actions to those of “children of a similar age” or development, instead of that child at his baseline, risks misidentifying the child’s usual behavior as a symptom of a manic episode. By comparing each child only to himself, symptoms of a developmental disorder or of a chronic disruptive disorder such as ADHD, oppositional defiant disorder (ODD), and conduct disorder, will not be confused with symptoms of a manic episode. When the presentation and history are too unclear to reach a decision, a longer period of observation, assessment, and monitoring is warranted.
It is important to consider the research on SMD in its historical context. In the 1990s, child psychiatry researchers suggested two “developmental modifications” to be instituted when diagnosing BD in children. One influential modification held that severe, non-episodic irritability was the defining feature of pediatric BD.13–15 According to this view, BD in youth typically presents, not as distinct manic episodes, but as non-episodic, severe irritability and symptoms of ADHD. Another influential modification recognized the importance of episodes but proposed that manic episodes in pediatric BD typically persist for years while mood cycles as short as four hours are characteristic of the illness.16–18 Such long-duration manic episodes seldom occur in adults with BD in modern times and, in children, are difficult to reconcile with the notion of an episode as a “distinct period.” Indeed, in this modified formulation, the length of a single episode was often a significant portion of the child’s life.17, 18 Thus, both of these influential views called attention to children with severe and persistent forms of mood disturbance without clearly demarcated mood episodes.
When these modifications were being advanced, researchers elected to exclude or include chronically irritable youth in their samples of patients with BD, but no research focused specifically on this phenotype. To facilitate research on chronic irritability, the SMD criteria (Table 1) were developed by Leibenluft et al. (2003), in consultation with investigators at other sites.9 The development of the SMD criteria made it possible for investigators to compare youth with a phenotype displaying very impairing, chronic irritability but lacking distinct manic episodes, to those with the classic, episodic presentation of BD. This allowed investigators to test whether the two clinical presentations are manifestations of the same illness.1 Importantly, there was no claim that the SMD criteria define a new diagnostic category. Instead, the criteria were designed to address a specific scientific question: to what degree do children meeting these criteria resemble youth with BD on variables used to validate psychiatric disorders?
It is important to recognize that children who present with chronic irritability (but no distinct manic episodes) should be differentiated from both those with BD and those with BP-NOS. This is important because chronic irritability is very common in children, far more common than BD,3, 19 and data suggest that children with chronic irritability differ from children with BD in outcome, family history, and pathophysiology. In studies at the National Institute of Mental Health (NIMH) Intramural Research Program, approximately 60% of youth meeting criteria for SMD were carrying a community diagnosis of BD at the time that they were referred for assessment, although none had a history of a distinct manic episode. Frequently, the rationale given for the BD diagnosis in these children is their very severe irritability.
Despite its high frequency and considerable morbidity across the age span, irritability is not defined explicitly in DSM-IV and it is included as a criterion for many disorders, including major depressive disorder (MDD), ODD, and generalized anxiety disorder (GAD). The definition of irritability in SMD has two components—one tonic and the other phasic.1, 9 The tonic emotional component consists of negatively valenced (i.e., angry/irritable or sad) mood. The tonic component is a persistent, incessant angry mood that presents between outbursts of verbal or physical aggression. The phasic component consists of outbursts that reflect increased reactivity to negative emotional stimuli. The outbursts are manifest by heightened verbal and/or physical aggression, typically in response to frustration. These outbursts are clearly excessive for the situation and must occur at least three times a week. Developmental considerations are important i.e., given the provocation, the outbursts must be beyond what would be expected of a child at that developmental level. The tonic component of irritability must be present most of the day, most days. For segments of the day, the child’s level of annoyance might be low, or even absent, but it should be a regular feature of the child’s mood “most” of the time i.e., more than half of the child’s waking hours.
In order to demarcate SMD from the developmentally normative increase in chronic irritability that reaches its peak around age 12 or 13,20 the criteria require that the syndrome have its onset before age 12. Also, because the SMD phenotype was designed to capture chronic irritability, both the tonic and phasic manifestations must be present for at least a year, with no symptom-free periods longer than 2 months. In SMD, symptoms may decline during summer breaks, but they do not vanish, distinguishing SMD from situational stressors such as poor adjustment to school and from BP-NOS, where distinct episodes (see below) are the defining feature. Finally, chronicity and severity are bolstered by requiring that symptoms extend beyond one situation (e.g. not only home, or with only one teacher); impairment from irritability must be severe in at least one setting and at least mild in one other among home, peers, and school.
SMD criteria also require “hyperarousal” symptoms, defined as at least three of the following: insomnia, agitation, distractibility, pressured speech, flight of ideas/racing thoughts, and intrusiveness. Hyperarousal symptoms were included because the SMD criteria were designed to capture those youth with ADHD plus chronic irritability who some contended should receive the diagnosis of BD.13, 14 Specifically, the hyperarousal criteria of SMD include those symptoms that overlap between the “B” criteria of a manic episode and ADHD. However, as noted above, in the case of a manic episode, the hyperarousal symptoms must be episodic, meaning that they occur in distinct periods of time during which there is also a definite change in mood. In contrast, in SMD, these symptoms, like all SMD symptoms, must be chronic, without marked fluctuations. As noted below, the most significant difference between DMDD, proposed for inclusion in DSM-5, and SMD is that the former does not include these hyperarousal symptoms.
Most youth who meet criteria for SMD also meet DSM-IV criteria for ODD and ADHD. This overlap occurs almost by definition, since the SMD criteria were designed to capture youth with severe irritability and ADHD symptoms whose nosological status with respect to BD was in doubt. However, there are important differences between SMD and ODD. First, oppositional as well as irritable symptoms are necessary in order to meet DSM-IV criteria for ODD. Second, roughly 85% of children with SMD fulfill criteria for ODD, but only approximately 15% of youth with ODD meet criteria for SMD, because SMD requires more severe and pervasive irritability than does ODD. Thus, SMD could be viewed as an extreme form of the irritable component of ODD.1, 21, 22
As with ODD, most youth with SMD meet DSM-IV criteria for ADHD. As noted above, because of the research question that the SMD criteria were designed to address, the criteria include those symptoms that overlap between the “B” criteria of a manic episode and ADHD. However, while the hallmark symptom of SMD is affective (i.e., severe irritability), the DSM-IV criteria for ADHD do not include emotional symptoms. Since SMD can be viewed as a particularly irritable subtype of ADHD, research on SMD parallels emerging interest in “emotional impulsivity” and related constructs in ADHD.23, 24
When the criteria for SMD were operationalized, exclusionary criteria were set so that the SMD phenotype would include those youth whose nosologic status with regard to BD was most controversial. Thus, one of the most important exclusion criteria is that the presence of any illness on the BD spectrum is exclusionary for SMD. A past or current history of a manic or hypomanic episode, even an episode that is as short as one day and therefore does not meet DSM-IV-TR for BD, excludes a subject from SMD.9 This excludes most youth meeting criteria for BP-NOS, and all meeting DSM-IV-TR criteria for Cyclothymic Disorder. The exception to this would be the approximately 20% of children with BP-NOS who have had more than four hours of symptoms in a day (see below), but less than the “most of the day” that SMD criteria require for symptoms to be considered an episode. Of note, the assessment techniques for SMD do not require ascertaining symptoms by the hour whereas, for BP-NOS, such assessment is important because the four hours needed for an episode do not need to be consecutive. These differences in assessment highlight the need for a study comparing youth who meet criteria for SMD with those meeting BP-NOS.
Other SMD exclusion criteria were designed to exclude youths whose irritability was clearly associated with a DSM-IV depressive or anxiety disorder. Thus, youth whose irritability was present only in the context of a major depressive disorder are excluded. With regard to dysthymic disorder (DD), only 13 of 183 (7%) children with SMD exhibit the chronic neuro-vegetative symptoms (sleep, appetite, and energy), hopelessness, or low self-esteem required for the diagnosis. With regard to anxiety disorders, youth with separation anxiety disorder whose irritability occurs exclusively in the context of separation, as well as those whose irritability is attributable to post-traumatic stress disorder, are excluded from SMD samples. In addition, given the complex issues that arise when any psychiatric illness occurs in the setting of a pervasive developmental disorder, such youth are also excluded from SMD.12
As detailed below and elsewhere,1 research comparing longitudinal course, family history, and brain function in youth provides strong, albeit still suggestive, evidence that SMD is not a pediatric phenotype of BD. However, while the research suggests that SMD is not a form of BD, it leaves open many questions as to what SMD is. As noted above, by definition, the overwhelming majority of youth with SMD meet DSM-IV criteria for ODD, and most also meet criteria for ADHD.1 It is interesting to note that only a small minority of youth with SMD meet lifetime criteria for major depression, despite the fact that their chief presenting complaint is irritable mood; however, approximately half meet criteria for a current anxiety disorder (e.g. generalized anxiety disorder [GAD], separation anxiety disorder [SAD], and/or social phobia) (Table 2). The aim of defining SMD was not to establish a condition that excluded all other diagnoses.
The same questions about diagnosing youth with chronic irritability that led to the definition of SMD motivated the DSM-5 Work Group to examine the classification of severely irritable children. The Work Group’s deliberations led to the proposal that a new diagnosis, i.e., DMDD, be included in the DSM-5.25 Importantly, the criteria for DMDD differ from those for SMD in that the former does not require “hyperarousal” symptoms; youth who also meet criteria for ADHD could receive both diagnoses. In addition, whereas youth with SMD must have the onset of their symptoms before age 12, the corresponding age for DMDD would be 10.25 Because of these differences in criteria, data on SMD cannot be assumed to generalize to DMDD, although the central clinical feature of severe, non-episodic, impairing irritability is common to both syndromes.
Given the relatively limited research on SMD and, indeed, on clinically significant irritability in general,1 it is not surprising that considerable controversy surrounds the question of whether DMDD should be incorporated as a separate diagnosis into the DSM-5.22 An alternative proposal is that a specifier should be created that operationalizes severe irritability and can be applied across a number of diagnoses. Valid arguments exist on both sides. Some have argued that it is preferable for very irritable youth to continue to receive multiple diagnoses, for example, ADHD, ODD, and mood disorder, perhaps with a specifier for severe irritability, thus recognizing the co-occurrence of these complex disorders. This has the advantage of closely mirroring clinical presentation and avoiding the implication of a unitary chronic irritability syndrome. This option was considered seriously. Ultimately, the decision as to whether to diagnose complex co-occurring disorders or a unitary construct was arrived at after careful consideration of the advantages and disadvantages of both options. Given the continued, but lengthy, search for biomarkers on which to base psychiatric diagnoses, the issue is unlikely to be resolved for many years.
Those favoring the establishment of a new unitary diagnosis express concern about the increased rate at which children presenting with chronic, severe irritability are inappropriately receiving the diagnosis of BD; note that the diagnoses of ADHD, ODD, and an anxiety disorder are typically insufficient to justify the degree of services required by these very severely impaired youth; argue on the basis of clinical experience that specifiers are often overlooked; and believe that this condition is a mood disorder, rather than solely a disruptive behavior disorder. Clinicians and researchers holding this perspective maintain that a unitary diagnosis will address these issues by creating a diagnosis, defined so as to be quite severe, that will be housed in the mood disorder section of the DSM. Those favoring a specifier note that there is significant overlap between SMD and ODD; there are as yet no specific treatments for SMD; the relevant research is limited and mostly from one research group; and the introduction of the SMD diagnosis may cause clinicians to overlook other diagnoses that might cause irritability and for which there are demonstrated treatments (e.g., major depression, anxiety disorders). Clearly, both positions have considerable merit; as noted above, the debate highlights the limitations of psychiatric diagnosis grounded entirely in symptom description without brain-based measures, and the problems of categorical vs. multi-dimensional approaches to psychiatric nosology.
Indeed, the limitations of categorical diagnoses have received considerable attention recently through NIMH efforts to develop Research Domain Criteria (RDoC).26 “Frustrative non-reward” is an RDoC construct denoting an environmental context that might elicit the clinical symptom of irritability. Like other RDoC constructs, frustrative non-reward can be studied in both humans and animals to advance pathophysiological and treatment research. Such research, coupled with continued longitudinal and genetic studies in humans,27 will enable the development of novel interventions for children impaired by chronic irritability.
Subthreshold presentations of bipolar disorder have long been described in both adults28–31 and youth.32–34 Over the last two decades studies have shown that subthreshold BD symptoms have prognostic and epidemiological implications that are important in understanding BD28–34,35. These subthreshold conditions, BP-NOS and Cyclothymic Disorder, are often combined with type I and type II BD into a “BD spectrum.” While the DSM criteria for a manic or hypomanic episode are operationalized and delineated, those for BP-NOS and cyclothymic disorder are quite vague.8 Although these subthreshold presentations can be associated with significant impairment, 34 definitions of subthreshold BD have varied widely among authors and the criteria were not always clearly defined (see Table 1 in Axelson et al., 2011)2. This contributed to widely divergent findings2. Since one important focus of the COBY study was subthreshold presentations of BD in youth, a clear definition was required. Taking into account the existing literature at the time and conventions for the assessment of BD in the Kiddie–Schedule for Affective Disorders and Schizophrenia (KSADS), the COBY study operationalized criteria for BP-NOS as shown in Table 3.
At the time that the COBY BP-NOS criteria were constructed, published studies defined significant subthreshold manic symptomatology in different ways, ranging from meeting all criteria for a hypomanic episode except for duration (allowing 1–3 day episodes)30 to the presence of only the A criterion (distinct period of abnormally elevated or irritable mood) without any duration criteria.34 The COBY investigators decided that the “one symptom short” and/or “short duration” was a reasonable compromise so that youth would have a reasonably high likelihood of having a bipolar illness, but be broad enough to include a substantial proportion of youth who present with symptoms that raised the possibility of a bipolar diagnosis. The threshold for the duration of symptoms in a given day (4 hour minimum) was taken directly from the Kiddie–Schedule for Affective Disorders and Schizophrenia–Present (KSADS-P) mania section elated mood item, that in turn was adapted from the Schedule for Affective Disorders and Schizophrenia (SADS) elation item which used a threshold of “several hours”.36,37 The 4 lifetime days criterion was established to ensure that an isolated day or two was not labeled as BP-NOS.
Thus, COBY-defined BP-NOS includes youth who are one B criterion short, and/or do not meet the duration criterion, for DSM-IV manic, mixed or hypomanic episodes. In addition, youth who have met DSM-IV criteria for a hypomanic episode but not for a major depressive episode would also meet BP-NOS criteria. DSM-IV criteria for cyclothymic disorder are so ambiguous that youth who might have been diagnosed with cyclothymia would generally meet BP-NOS criteria and, thus, they were included in the BP-NOS category. Some investigators note that in clinical application, DSM-IV criteria can make it difficult to distinguish BP-NOS from cyclothymic disorder in youth.38 Whether it is valid to combine these subtypes will require studies that apply precise definitions and compare them.39 When applied to the sample of youth referred to the COBY study with possible BD, 35% of the 438 subjects met criteria at intake for BP-NOS (58% were BD-I and 7% were BD-II).7 About 85% of the BP-NOS subjects met or exceeded the number of B criterion symptoms for mania/hypomania, but did not meet the episode duration criteria, while 12% had a history of a hypomanic episode with a major depressive episode, and only 3% did not meet the B symptom criteria.
Data to date indicate that BP-NOS may be a phenotype on a diagnostic continuum with BD-I and II. At the intake assessment, BP-NOS and BD-I subjects did not differ in age of onset, duration of illness, lifetime rates of comorbid diagnoses, suicidal ideation or major depression, family history, or the specific manic symptoms that were present during the most serious lifetime episode. Though irritable mood was a common feature of the most severe episode of manic symptomatology for both BP-NOS (80%) and BP-I (85%) subjects, both groups also had high rates of elated or expansive mood (BP-NOS 82%, BD-I 92%). However, BP-NOS differed from BD-I subjects in features associated with illness severity. On average, BD-I subjects had more severe manic symptoms and greater overall functional impairment than youth with BP-NOS, as well as higher rates of hospitalization, psychosis, and suicide attempts.
Over longitudinal follow-up, BP-NOS subjects continued to have significant mood symptomatology and were euthymic for less than 40% of the weeks during follow-up. 40 Compared to youth with BD-I, those with BP-NOS took longer to recover from their index mood episode but, once in remission, had a lower hazard for recurrence.40 Over an average of five years follow-up, 45% of BP-NOS youth progressed to a diagnosis of BD-I (23%) or BD-II (22%). Amongst the variables measured at intake assessment, the main predictor of diagnostic progression was a history of mania or hypomania in a first or second degree relative.2 Most (76%) youth with BP-NOS had at least one DSM-IV mood episode over follow-up, including 49% with a manic, mixed or hypomanic episode and 57% with a major depressive episode. 2 Over prospective follow-up, youth with BP-NOS had the same rate of suicide attempts as those with BD-I and BD-II 41 Thus, taken together, these data indicate that BP-NOS youth continue to have a high burden of mood symptoms over prospective follow-up and many progressed to clear BD.
It is critical to note that the BP-NOS criteria were intentionally broad, as these were research criteria designed to study youth who may or may not have bipolar spectrum illness. However, the interpretation of these BP-NOS findings must consider that this was a referred clinical sample, so the subjects likely represent youth whose symptoms are particularly severe. Thus, although the BP-NOS criteria were broad, nearly all of the BP-NOS subjects met full symptom criteria for mania/hypomania, and most far exceeded the criterion requiring four symptomatic days lifetime. A sub-analysis of BP-NOS subjects who had quantifiable lifetime duration data available, and who met BP-NOS criteria because of insufficient episode duration, indicates that 94% had more than 10 estimated total number of lifetime days meeting BP-NOS criteria, and 68% had more than 50 lifetime days. In addition, in 54% of these subjects, the estimated longest lifetime episode of manic symptoms was between 2 and 4 days (David Axelson, personal correspondence 2012).
Though these data are a first step toward establishing the validity of BP-NOS as a diagnostic category in youth, replication is required. To explore the validity of BP-NOS criteria further, and to learn whether they may be prognostically informative and clinically relevant beyond the original cohort, they were used in both the Longitudinal Assessment of Manic Symptoms (LAMS) study and the Pittsburgh Bipolar Offspring Study (BIOS).4,40 The LAMS study seeks to identify the longitudinal course of youth presenting to outpatient clinics with “Elevated Symptoms of Mania,” (ESM), defined by high scores on a screening instrument called the Parent General Behavior Inventory 10-item Mania Scale (PGBI-10M). Fourteen percent of youth who screened positive on the questionnaire met BP-NOS criteria, while 11% met criteria for BD-I and 1% for BD-II.4 The BIOS, a study of offspring of adults with BD, found that, at intake, offspring of parents with BD were 13 times more likely to meet BP-NOS criteria than demographically matched community controls (7.2% vs. 0.4%).40 Furthermore, in the offspring of BD parents, BP-NOS was much more common than BD-I (2.1%) or BD-II (1.3%). Of course, more remains to be learned before BP-NOS criteria can be applied in a clinically meaningful way. Longitudinal follow-up data from the LAMS and BIOS samples will provide additional data regarding the validity of the BP-NOS phenotype. While these data are being gathered, a growing literature suggests that adults with major depression and recurrent, subthreshold duration (1–3 day) hypomanic episodes resemble adults with full-duration hypomania in rates of bipolar family history, age of onset of depression, and symptoms of atypical depression.42–44 Taken together, these studies can advance our understanding of the implications of subthreshold hypomanic symptoms across development.
As noted above, with the exception of a family history of bipolarity, prospective data did not provide evidence for phenotypic features that predicted diagnostic progression from BP-NOS to BD-I/II. Thus, these data do not provide an empirical basis for revising the BP-NOS criteria so that they predict conversion to BD more consistently. Considerable controversy has resulted from the fact that the lower bound of the BP-NOS criteria is defined as four symptomatic days; each day must include at least four hours of symptoms; and the four hours need not be consecutive. Of note, similar controversy also exists as to the appropriate lower bound duration for a hypomanic episode in adults.42, 45 The BP-NOS definition was designed to operationalize a minimum time threshold for distinct (hypo)manic symptomatology and to ensure that such symptomatology be present for a significant part of the day, given that youth with significant subthreshold manic symptoms are usually not observed all day by adult informants. In addition, these youth often have subthreshold depressive symptoms and irritability, rather than clear manic symptoms, during much of the day, and there can be diurnal variation in mood state. It can be difficult to establish duration of manic symptoms, and caregivers and youth can both over- and underestimate how long these periods of mania last.
When applying COBY BP-NOS criteria to a general clinical population, one should consider four pertinent facts. First, all COBY participants had been pre-screened and had a suspected bipolar diagnosis. Therefore, youth in a clinic population who meet COBY BP-NOS criteria (Table 3) but have not been pre-selected for probable bipolar disorder may not exhibit similar severity, course, and outcome to those in the COBY study. Second, a patient can meet COBY BP-NOS criteria by virtue of episodes that 1) are too short to meet the criteria for hypomania, OR 2) fall one symptom short of the number of symptoms required for a hypomanic episode. However, in practice, nearly all COBY BP-NOS subjects had episodes that included the requisite number of symptoms for hypomania, but were too short in duration. Third, in the majority of COBY BP-NOS subjects, episode duration, while shorter than that required for hypomania, exceeded the minimum duration specified in the COBY criteria. Lastly, nearly all COBY BP-NOS subjects had many recurrent short episodes that, together, well exceeded the cumulative requirement of four days lifetime.
Given these facts, there is concern that, if the original COBY BP-NOS criteria were applied to a general clinical population, they would ascertain patients who are considerably more heterogeneous than those described in the COBY study, and therefore that findings from the COBY data may not generalize to these youth. Hence, we recommend the following modified criteria for a diagnosis of BP-NOS in a general clinical setting.
The presence of (1) or (2):
These criteria are more restrictive than the original COBY BP-NOS criteria, because they require the same number of symptoms as hypomania and that symptoms persist for most of a day, rather than for a minimum of four hours. We believe that the new definition will be easier to use and reduce the possibility that multiple temper tantrums in a day, or rebound irritability and ADHD symptoms from stimulants wearing off, would be labeled BP-NOS. However, it is certainly possible that some youth with a bipolar illness will not meet these modified BP-NOS criteria. Some COBY subjects experienced brief, recurrent manic/hypomanic symptoms lasting only a few hours at a time and yet progressed to bipolar I disorder. In general, these brief episodes were not described as extended temper outbursts, but had features specific to mania, albeit at much shorter durations. Also, these brief periods were often embedded in, or alternated with, periods of significant depression and/or irritability. Studies should compare patients meeting the original COBY BP-NOS criteria versus the modified ones proposed here. We believe that, relative to the original criteria, the modified criteria will be easier to apply and will balance specificity and sensitivity more optimally for a broad clinical population.
In addition to the clinical differences between BP-NOS and SMD (e.g., presence of mania/hypomanic symptoms and episodicity), two other important lines of evidence argue for making a distinction between BP-NOS and SMD (Table 4). Specifically, when comparing cohorts of children with BP-NOS and SMD, evidence suggests different rates of conversion to BD and different prevalence of parental BD.
In the BP-NOS studies, at two years,6 four years,46 and five years2 of follow-up, the rate of conversion from BP-NOS to BD was 25%, 38%, and 45% respectively. This stands in stark contrast to the rate of 1.2% that was seen in children with SMD47, 48 after 29 months of follow-up, and to the rate of 0% from a mean 8-year epidemiological follow-up study.3 In addition, epidemiologic data suggest that chronic irritability in adolescence does not predict BD in adulthood, but instead predicts increased risk for unipolar depressive and anxiety disorders.20, 47 These data are consistent with longitudinal studies indicating that the irritable dimension of ODD predicts subsequent depression and anxiety, whereas the headstrong dimension predicts antisocial outcomes.48–51 Of note, other data52, 53 suggest that children previously diagnosed with ODD have an elevated risk for almost all psychiatric disorders, including BD, in early adulthood; these studies do not separate predictions from the irritability vs. headstrong dimensions of ODD.
In addition to differing in longitudinal course, youth with SMD and those with BP-NOS appear to be dissimilar in their family history. Specifically, rates of a family history of BD are considerably higher in children with BP-NOS compared to those with SMD. In the BP-NOS studies, roughly 25% of those with BP-NOS had a first-degree relative with BD, a rate virtually equal to that seen in the first-degree relatives of those with BD.6, 54 In a pilot study of children with SMD, however, the rate of a parent with BD was 3%,55 which approximates the general population prevalence.56, 57 Congruent with the conversion data, these family history data suggest that BP-NOS is on a pathophysiological continuum with BD, whereas SMD is not.
Unfortunately, clinicians have little specific treatment literature to guide them in the treatment of either SMD or BD-NOS. In the only published controlled clinical trial of SMD, which included a small sample, the response to lithium carbonate and placebo was comparable.58 There are no blinded, controlled clinical trials in children with BP-NOS, and the evidence for efficacy of lithium in pediatric BD has been disappointing.17, 59 Therefore, it is impossible to say whether the same psychopharmacologic and psychotherapeutic interventions should be used for SMD and BP-NOS, or for these disorders and BD. Nonetheless, differences between BP-NOS and SMD in longitudinal outcome and family history suggest that these conditions may be dissimilar in their underlying pathophysiology, and thus one cannot assume that they should be treated according to the same guidelines. Controlled trials in each of these populations are imperative.
Given the lack of efficacy of lithium, and the unclear nosologic status of SMD, clinicians treating a youth with SMD should first identify DSM-IV diagnoses for which the youth qualifies and use treatments that have demonstrated effectiveness in those disorders. For example, if a youth with SMD meets criteria for ADHD, stimulant treatment would be indicated; if a youth with SMD meets criteria for an anxiety disorder or MDD, cognitive-behavioral therapy (CBT) and/or selective serotonin reuptake inhibitor (SSRI) treatment would be indicated. Prior research suggests that, when irritability is a component of the presentation of one of these disorders, it improves during the treatment of the disorder. In contrast to these guidelines, if a child with chronic irritability who meets criteria for SMD, ADHD, and anxiety is viewed as having BD and is treated as such, then treatment with only stimulant and SSRI would be relatively contraindicated. Instead, a second-generation antipsychotic (SGA) or mood stabilizer might be considered first-line treatment for BD, with the possible addition of stimulant and/or SSRI as indicated.59 This is one example when differentiating between symptoms of SMD and BD can be important.
Indeed, many clinicians who associate severe, non-episodic irritability with BD are reluctant to prescribe stimulants and/or SRIs to youth exhibiting the SMD phenotype. However, given that many children with SMD meet criteria for anxiety disorders and ADHD; that longitudinal data show associations among SMD, anxiety, and unipolar depressive disorders; and that data support the efficacy of stimulants and, to a lesser degree, SSRIs in the treatment of irritability, these two classes of medications could play a prominent role in the treatment of SMD. Specifically, two meta-analyses demonstrate moderate to large effect sizes for stimulant treatment in decreasing aggression, a common symptom of irritability, in youth with ADHD.60, 61 Treatment trials in adults suggest that SSRIs can be beneficial in the treatment of irritability, aggression, and/or explosive outbursts that occur in the context of depression, dysthymic disorder, or anxiety.62, 63 Pilot data from placebo-controlled trials, while not definitive, could provide some information on the safety and efficacy of SSRIs for children with chronic irritability. In addition, there are no data on efficacy or safety of non-stimulant medications, such as atomoxetine or alpha-adrenergic agonists, in SMD.
The most studied agents for aggression and irritability in children are SGAs, which are considered first-line pharmacological interventions in some reviews.64 Several studies have reported SGAs were effective in reducing impairing irritability and aggression in children with sub-average IQ,65 in children with ASDs66–68 and, most relevant to SMD, those with disruptive disorders.69–71 In addition, there is older literature on haloperidol for aggression and irritability in children.72 Unfortunately, as well documented elsewhere,73 antipsychotics carry a high risk of metabolic and cardiac side effects.
A few studies have evaluated the efficacy of mood stabilizers such as divalproex or lithium in the treatment of aggression. A small pilot study in aggressive children with ADHD, who were non-responsive to stimulants, found that divalproex, in combination with stimulants and behavior therapy, was more effective than placebo, stimulant, and behavior therapy.74 Campbell et al.72, 75, 76 found that lithium was useful in conduct disordered aggressive youth, but other studies77 did not. And, as mentioned above, a controlled trial of lithium vs. placebo in SMD did not provide evidence for efficacy.58
Clearly, there is an urgent need for more clinical trials in youth with severe irritability. Also, if future pediatric clinical treatment trials for anxiety, depression, ODD, or ADHD collected continuous measures of irritability, investigators could conduct post-hoc analyses of treatment effects on irritable mood. In addition to identifying efficacious psychopharmacologic and psychotherapeutic interventions, such trials could also elucidate how severe irritability in the context of ADHD or anxiety influences treatment response.
Treatment guidelines for BP-NOS do not exist. One small, double-blind, placebo-controlled study of youth who had a parent with bipolar illness and were diagnosed with BP-NOS or cyclothymia found no effect of treatment with divalproex sodium.78. A controlled study that included some adolescents with BP-NOS demonstrated modest benefit of a family-focused psychoeducational treatment for adolescents with “bipolar spectrum illness”, but results for the BP-NOS subgroup were not reported.79 Another controlled study for children with mood disorders included some children with BP-NOS and reported benefits from a group family psychoeducational treatment; however, results were not reported for separate diagnostic subgroups.80
Without controlled trial data, treatment recommendations cannot be based on scientific study. For youth with BP-NOS, if the situation is not clinically urgent, psychosocial treatments, if available, may be the best option. However, when psychosocial treatments are not available, the response is inadequate, and/or functional impairment is high, then pharmacotherapy is usually necessary. Most research participants with BP-NOS received pharmacotherapy; on average, they were exposed to antimanic medications for about 50% of the longitudinal observation period.2 SGAs appear to have the clearest evidence for efficacy in treating acute manic episodes in youth,59 so they are a logical choice for youth with BP-NOS who require pharmacotherapy. However, an ongoing evaluation of the potential risks and observed benefits is necessary, particularly when a bipolar diagnosis is not clear. Other anti-manic agents such as lithium or divalproex can be used if SGAs do not address manic symptomatology adequately or are not tolerated. Stimulants can be considered for comorbid ADHD if anti-manic agents have addressed manic symptoms successfully.81 Depression and anxiety can be managed using psychotherapy (e.g., CBT) as first line treatment. Refractory depression and anxiety can be treated with SSRIs. While it is important to observe for side effects and exacerbation of mood symptoms, two recent reviews both concluded that the risk of inducing a switch into mania with SSRI treatment has not been well-quantified and may be relatively low.82, 83
Treatment decisions are difficult when a child or adolescent has a history of mild, transient and relatively non-impairing episodes meeting BP-NOS criteria, but is presenting with impairing symptoms of depression or another disorder, such as ADHD or an anxiety disorder. If an appropriate psychosocial treatment for the target symptoms (e.g. CBT for anxiety or parent management training for ODD) is available, it should be tried first. If pharmacotherapy is required, then it is important to have a frank discussion with patient and caregivers about the diagnostic issues and the potential risks and benefits of different treatment options. Reviewing the adequacy of, and responses (negative and positive) to, prior medication trials provides useful information. In many cases, it may be reasonable to initiate treatment with an SSRI or stimulant cautiously, as long as the patient and family are counseled to seek assistance immediately if negative effects on mood and/or behavior occur. However, factors such as a clear, strong family history of bipolar disorder, history of prominent episodes meeting BP-NOS criteria, and poor response to multiple adequate medication trials in the past, may in some cases justify a clinical judgment to use an SGA first and then add on other medications, depending on clinical response.
Further research is needed to identify effective treatments for these very impairing syndromes. In addition, it is important to try to identify clinical variables and biomarkers that predict conversion from BP-NOS to BD-I or II, and from SMD to anxiety or depressive disorders. More study is also needed of subclinical presentations of BD and of the relationship of SMD to other mood disorders and anxiety disorders.
A key question is how the neural circuitry mediating the symptomatology of BP-NOS compares to that of children with BD and SMD. Data show differences between BD-I or II and SMD in the neural dysfunction mediating deficits in face-emotion processing84 or cognitive flexibility85 and in the recruitment of attentional resources during frustrating tasks.86, 87 For example, on a face-emotion task, youth with SMD and BD reported similar subjective ratings of fear and hostility to faces, and displayed similar reaction times; however functional magnetic resonance imaging (fMRI) data revealed that those with SMD showed hypoactivation in the left amygdala when rating their fear of neutral faces, compared to both BD and healthy volunteer children.84 In contrast, children with ADHD who did not meet irritability criteria showed increased activation in this region compared to youth with BD and to healthy controls.84 Using tasks that require cognitive flexibility, behavioral performance is compromised in both youth with SMD and those with BD. However, fMRI measures during incorrect trials in a response reversal task found that, compared to youth with BD, those with SMD showed significantly less activation in the inferior frontal gyrus. 85 In BP-NOS, a recent fMRI study demonstrated differential patterns of abnormal neural activity in, and connectivity between, neural regions supporting emotion processing and regulation in BP-NOS versus BD-I youth and healthy subjects.88 A productive line of research might compare neural function among youth with BD-I, BP-NOS, and SMD. One might predict that, on face emotion processing or frustration tasks, neural dysfunction in BP-NOS would differ more from that seen in SMD than from that seen in BD-I.
Taken together, the current literature suggests that being mindful of a distinction between BP-NOS and SMD is important for clinical care and research. The prevalence and severity of these conditions make it likely that practicing clinicians will encounter patients with these symptoms. Without a clear designation in the DSM, these phenotypes pose a nosological challenge to clinicians who are pressed to offer appropriate diagnoses to patients and the families who care for them. Future work comparing these populations is needed.
This research was supported in part by the Intramural Program of the National Institute of Health, NIMH. The authors thank Ms. Kendra Hinton of NIMH for her assistance in the preparation of this manuscript.
Disclosure: Dr. Birmaher has received research funding from NIMH. He has received or will receive royalties for publications from Random House, Inc. (New Hope for Children and Teens With Bipolar Disorder) and Lippincott Williams and Wilkins (Treating Child and Adolescent Depression). Drs. Towbin, Axelson, and Leibenluft report no biomedical financial interests or potential conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Dr. Kenneth Towbin, Emotion and Development Branch, Intramural Research Program, National Institute of Mental Health (NIMH)
Dr. David Axelson, University of Pittsburgh and University of Pittsburgh Medical Center/Western Psychiatric Institute and Clinic.
Dr. Ellen Leibenluft, Emotion and Development Branch, Intramural Research Program, National Institute of Mental Health (NIMH)
Dr. Boris Birmaher, University of Pittsburgh and University of Pittsburgh Medical Center/Western Psychiatric Institute and Clinic.