A commentary on youth onset bipolar disorder

A half-century ago, it was thought that children could not have a diagnosis of major depressive disorder (MDD). With the acceptance of pediatric-onset MDD as a diagnosis, we now know more about the origins, trajectories, and treatments for MDD. Without any diagnostic recognition, we could not have treated pediatric MDD as we do now nor could we design appropriate research studies. Indeed, the 30-year-old doubts about the existence of pediatric MDD or American exceptionalism about depression prevalence among youth are virtually nonexistent today.

There appears to be lingering uncertainty about the diagnosis and trajectory of pediatric bipolar disorder (PBD) as illustrated by recent position papers. Without any diagnostic recognition, research into the evolution of PBD, its trajectories, and its potential therapeutic interventions would be stymied. How could concerns about the evolution toward mania (or not mania) be evaluated without thorough and timely evaluation of its precursors? Indeed, the risk-to-benefit ratio of not accepting the validity of a diagnosis is much higher than the risk of holding it as a working hypothesis. Moreover, it may prevent the treatment of youth with highly disabling and even life-threatening symptoms. Here, we review a few critical findings to recapitulate what is known from research in PBD to date, which we believe to provide a helpful framework for pivotal next steps in the broader field of bipolar disorders.

1. Bipolar disorder is a lifelong condition that most frequently begins in childhood or adolescence. We must pay attention to signs of continuity between childhood, adolescent, and adult bipolar disorder (BD). With similarities in rates of hospitalization and impairment,¹ these characteristics point to a unified morbid process. Heart disease is an example of such a condition in which the scientific trend has been to understand precursors, early stages, and identify opportunities for earlier and less invasive intervention. Asserting that a lifelong chronic condition does not exist before a certain age, in contrast, would be more of an ostrich strategy, analogous to a fire department saying, “Don’t call us until you can send pictures of actual flames.” We appreciate and are aware of the concern about the challenges associated with diagnosing BD in younger and especially pre-pubertal children versus older adolescents, as well the potential pitfalls from the use of the term “paediatric” or “pediatric.” These diagnostic challenges signal a need for careful and comprehensive assessment, using advanced diagnostic methods. However, given the variability in ages of onset, it may be premature to designate age cut-off before which BD is not considered a valid diagnosis.

2. Secular trends indicate that globally, the onset of puberty is occurring earlier, which may contribute to observations of mood disorders in younger populations than previously considered.² The question then becomes whether it is conceivable that mania can occur prior to pubertal onset. If there is no question that mania can occur after puberty, why can it not occur before puberty? The literature is overwhelming that PBD has a more pernicious course than adult-onset illness, consistent with naturalistic observations in clinical settings. More than a thousand pediatric-onset bipolar patients have been followed for many years at multiple academic settings across the United States, and a majority shows persistence of illness with symptoms occurring in about 80% of those followed-up. In most of these studies, the average age of onset is 11–12 years of age, indicating that many must have had a prepubertal onset. As an example, a large study of adults with BD (n = 3,658) showed that about 70% reported mood onset before age 18 of which 30% were before age 12. In these studies, comprehensive evaluations of parents and children was conducted, adhering to formal research diagnostic criteria.

How early mania can occur is an intriguing question that merits our curiosity (see illustrative vignette in the Supplement). There is accumulating body of evidence that is converging on genetic- and brain-based pathophysiologic models of BD in carefully characterized research cohorts that are ascertained quite differently from epidemiologic or hospital electronic medical record databases. Nevertheless, there is resounding agreement across all databases, retrospective or otherwise, that onset of symptoms often occurs during childhood. In broad brushstrokes, the interesting question is “how early can this emerge, and under what circumstances? What are the factors that could mitigate progression?” Asserting by fiat that below a certain age mania is verboten shuts down inquiry and ignores a large body of work that has accumulated to date. How old someone must be to have a “real” diagnosis of BD is probably no more answerable than defining a minimum age threshold for cancer or heart disease, putting most arguments questioning the validity of PBD into a dubious perspective.

3. Clinicians may be just as likely to miss mania (false negative) than to overcall it (false positive). One-third of youths with “unipolar” depression turn out to have BD. The average delay in diagnosing BD is 8–10 years and delays are longer for those with earlier ages of onset. Age of onset of BD in adults averaged 26 years in individuals without a family history of psychiatric illness or adversity in childhood. However, in those with a loaded family history and much adversity in childhood, the average age of onset was almost always below age 13. We try to remain sensitive to the evolution of mania because there is plenty of evidence that syndromal mania is preceded by a prodrome, which infrequently, may occur before puberty. Many youths who develop BD declaratively in adolescence present with subsyndromal mania as well as anxiety and ADHD comorbidities in childhood that are equally impairing. Such youth may not meet adult diagnostic criteria for BD and should not be diagnosed prematurely. Further, children (at any age but especially those who are prepubertal) with an only explosive temper, chronic mood lability, and disruptive behavior should be carefully evaluated before diagnosing them with PBD. Indeed, some children, particularly those with severe mood dysregulation, may be misdiagnosed with BD. These children may be treated with unnecessary medications and sometimes do not receive the appropriate treatment that can be helpful for more likely conditions (e.g., a very anxious child could be misdiagnosed with BD and consequently not offered treatment with SSRIs). There may also be youth misdiagnosed with BD who do not have the criteria needed to diagnose them correctly, but that this observation occurs does not mean that BD does not exist in children. Children and families benefit from repeated assessments to track symptom evolution, whether it is toward mania or not. We should iteratively ask ourselves whether the odds of missing a diagnosis of BD outweigh the odds of overcalling it.

We have seen a significant improvement in sensitivity and specificity for screening of BD in youth.³ As true for misdiagnosis, missing a PBD diagnosis can result in multiple failed and sometimes inappropriate treatments, iatrogenic induction of mania, and preclusion of progress toward effective novel and personalized therapeutics. Moreover, it is often the case that a PBD diagnosis is missed when common ambiguities present, such as heterogeneous clinical presentations, biases in self-report, sociodemographic differences, or clinician biases that BD cannot exist in children. These issues can be resolved with a correct diagnosis through a careful and systematic evaluation.

4. Bipolar disorder is a familial disorder that predisposes individuals toward developing a major mood disorder compared with the general population. Familial aggregation patterns seem to be distinct for PBD compared with pediatric MDD and other psychiatric disorders (e.g., ADHD), suggesting that the genetic wiring is different from the outset. Further, healthy offspring of parents with BD are neurophenotypically different than healthy offspring of parents with MDD or without any psychopathology. BD is likely polygenic and does not follow Mendelian genetics, so multifinality in outcomes for offspring is common. Despite variable outcomes, BD still has neurobiologic consequences that are likely impacting generations and merit further study.

A well-assessed family history, and in conjunction with polygenic risk scores and risk calculators, can provide helpful information to a practitioner to weigh the relative risk-to-benefit ratio for treatment planning, as evidenced by studies suggesting that familial BD has a differential response to antidepressant and lithium treatment compared with nonfamilial BD. Longitudinal analyses tell us about the continuity of childhood-onset BD into adulthood, but variable long-term outcomes indicate that we should be looking for clues to understand the personalization of BD interventions. Childhood is an optimal period to understand the roots and trajectories of BD given the ways that neurodevelopment introduces variability (including through spectrum conditions) and is independent of the confounding influences of long-term treatment or cooccurring conditions like substance use.

5. Despite opinions to the contrary, bipolar disorder has been observed in youth around the world, not just in the United States.⁴ Although rates for BD-I and BD-II are comparable around the world, rates of unspecified (previously not-otherwise-specified, NOS) BD are variable, possibly due to its lack of clear operational criteria, which may be contributing to ongoing debates about the existence of BD in youth, and merits further evaluation. In countries where PBD is relatively rare, clinicians will still need to consider the public health implications of delaying diagnosis and treatment. Where the diagnosis is more common, lack of study and accumulation of knowledge about course and treatment will have an adverse impact on the health and well-being of hundreds of thousands of children. Many who treat these children agree that they are highly dysfunctional and prone to acquire further impairing comorbidities (e.g., substance abuse) if left untreated. As true for all medical conditions, we as clinicians and researchers must balance the risk of a false positive against the risk of a false negative.

6. Discrete, episodic mania symptoms occur on a continuum/spectrum of severity and duration. A reliable diagnosis of bipolar I disorder hinges on specialized training, with consideration of developmental nuances and context before committing to the diagnosis. Nevertheless, the phenotype of classic mania is very specific to a diagnosis of BD, which can be reliably diagnosed with training. Training helps to appreciate the classic phenotype, but it is even more labor-intensive when clinicians attempt to understand spectrum conditions, including bipolar II, unspecified BD, for which often there are inconsistent definitions used particularly prepubertally, and cyclothymic disorder. With good assessment techniques, these subcategories can be just as reliably diagnosed in youth as bipolar I disorder. The point that community clinicians historically have not had much training in diagnosing PBD exposes an opportunity for improving care via better education and tools.

7. Periodicity and heterogeneity are defining features of bipolar disorder, which have not precluded phenotyping in adults. A common argument against the diagnosis of PBD is that youth with chronic irritability and explosive outbursts have attention deficit hyperactivity disorder or disruptive mood dysregulation disorder, not BD. Although this was more obvious in the past, currently, few modern conceptualizations of PBD include youth with chronic irritability. A considerable number of studies indicate that youth with BD have more exaggerated presentations of mixed mood and frequent shifts in mood polarity, often with high rates of comorbidity with other disorders. These phenotypic attributes do not invalidate the diagnosis of PBD but provide support for why a careful evaluation of accompanying symptoms and context is essential. Irritability is only classified as a symptom of PBD if it is episodic and present alongside elation and grandiosity, or other sentinel features of classic mania such as the decreased need for sleep or increased energy and activity, suggesting that it may have a different significance in PBD compared with other conditions. Even if irritability is nonspecific, it is still a symptom that should trigger a comprehensive evaluation, in the same way, that persistent headaches should be evaluated as indicative of anything ranging from allergies to migraines to a brain tumor.

8. Youth who are misdiagnosed or left undiagnosed tend to do poorly long term. Longitudinal data sets across the lifespan suggest significant morbidity in those youth with shifting symptom presentations. Unfortunately, individuals with BD also have a high prevalence of abuse, which interacts with more complex presentations of BD. Mania may define bipolar I disorder, but heterogenous depressive episodes and symptoms as well as mixed states can dominate the longitudinal course of BD. These are illness course facts that raise the importance of early detection and observation rather than negate them.

9. There are multiple evidence-based pharmacologic treatments that were studied in thousands of youths down to age 7 with PBD that have successfully obtained US regulatory approval, demonstrating large effect sizes for treatment compared with placebo. Lithium is the only US Federal Drug Agency (USFDA) approved medication for BD in youth down to age 7 based upon data from the Collaborative Lithium Trials. For youth with PBD, evidenced-based solutions may not have been possible without clear diagnostic criteria. It is also true that children without BD, but with high degrees of emotion dysregulation, irritability, or nonspecific mood changes respond equally well to similar treatments. However, the existence of this range of conditions is not precluded. The placebo effect is not large in pharmacologic trials of PBD (vs MDD), and multiple agents have USFDA approval for the treatment of mania in youth. In fact, the USFDA encourages that drugs approved for mania in adults also be studied in children, at minimum to confirm safety. Having multiple options for treatment opens opportunities for personalized treatment selection as patients and caregivers weigh relative risks and benefits.

Proactive early identification optimizes outcomes in medicine. For youth, these options include psychosocial interventions prior to or in concert with the initiation of a pharmacologic trial. In fact, if cases were detected earlier in childhood, psychosocial interventions like family focused therapy and interpersonal and social and rhythm therapy may not only be necessary but may play a role in prevention.⁵ Indeed, comparative effectiveness trials are needed for understanding the differential effects of evidence-based treatments on outcomes. Evidence generation in youth must be prioritized over nonevidence-based trial and error off-label strategies that abound and leave patients and families exposed to more risks in the long term, including death by suicide, recurrent hospitalizations, psychosis, as well as academic and treatment failures.

CONCLUSION

Society needs scholars to operationalize definitions of early onset BD, offer solutions to treatment dilemmas, and give hope to families coping with this condition. The field must make earnest efforts to communicate effectively about the nature and origins of BD and collaborate to solve the problems of delayed diagnosis and inappropriate treatments. Debates can help to advance science when there are no available answers to fundamental questions, but data are not lacking in PBD.

Questioning the validity of a diagnosis of BD among adults with confirmed BD who had their onsets in childhood benefits neither them nor their children, especially since early onset BD in a parent may be a risk factor for BD in offspring. It is high time to focus our energies to provide our early onset patients with systematic and developmentally sensitive approaches to diagnosis and evidence-based treatments and to continue iterating toward better solutions for one of the most complex and fascinating of all medical conditions.