In this update, I will discuss how bipolar disorder has been classically defined and the revisions to those definitions that are being proposed for DSM-5. Then, I will briefly discuss several of the treatment innovations made over the past 60 years, including treatment guidelines and second-generation antipsychotics (SGAs) used as mood stabilizers. Finally, I will touch upon a few innovations on the horizon for the treatment of people with bipolar disorder.

Bipolar I disorder

How do we define bipolar disorder (BD)? There is well-established historical and worldwide recognition of bipolar I disorder (BDI), the type of BD associated with manic episodes (1–3). Classically these episodes are not small perturbations of energy, mood, impulsivity, and action but, before the availability of treatment, manic episodes could lead to death due to physical exhaustion (2). The recurrent nature and severity of BDI is now better understood than in earlier times (1, 2, 4). After two clear episodes of mania, lifelong medication is recommended because of the greater than 95% probability that there will be a recurrence (2, 5). Mania alone (unipolar mania), is a relatively rare phenomenon accounting for less than 5% of all patients with BDI; and the majority of individuals with BDI also experience periods of depression.

Bipolar II disorder

There is now greater understanding and appreciation of the severity of BD type II (BDII), an illness with a predominance of depression, periods of mood lability, and some degree of hypomania (2, 3, 6).

Part of this increased awareness of BDII occurred as more medications became available for treatment of BD in the 1990s, and for the first time treatment guidelines began to emerge (7–9). The emergence of treatment guidelines brought clarity and focus to the essential absence of data on treatment management of patients with BDII in the late 1990s. Although seemingly it is not possible to develop guidelines without data, most early efforts, including those of APA (7–9) and the Canadian Network for Mood and Anxiety Treatments (CANMAT) (10, 11), recommended following the same guidelines for BDII as for BDI. In addition, formal studies of medication treatment of BDII are limited, and psychosocial therapies are rarely if ever directly compared with medications (2, 6, 12).

This distinct subset of patients with BDII was first clearly identified and validated a number of years ago (6, 13–15). Historically, in some sense BDII was seen as a “less severe” illness because of the absence of manic episodes. More recently, the severity, chronicity, impact on quality of life, and suicidal risk of this population have emerged in a number of studies supporting a different illness but of essentially equal severity as BDI (6, 15–18). In particular, the depression burden is significant and often chronic (19). Despite the improved recognition and studies of individuals with BDII, treatment studies have lagged behind. Debate continues as to the specificity of BDII disorder despite scientific efforts supporting reliability and validity (20).

Bipolar spectrum

Not as clear as BDI or BDII is the often stated “bipolar spectrum” or subthreshold bipolarity. When does a mild, less impairing, and less distressing degree of mood lability or depression become an illness appropriate to manage with medication and other tools?

The large group of patients currently classified as BD not otherwise specified (BD NOS) are poorly defined as a category. The current DSM-IV-TR provides limited guidance to specify individuals in this group, but rather refers to “examples include” (3). The important consideration of when symptoms become an illness or disorder must, in this group, be assessed not only by the presence or absence of a symptom but also by the degree of distress and impact on function.

These definitions are important because they provide a way to define a group, be specific in treatment approaches, and work to understand the underlying causes of the cluster(s) of symptoms defined as BD. In the new DSM-5, efforts are underway to improve the specificity and sensitivity of the diagnosis of BD and related diagnoses.

Depending on which epidemiology study is examined and how “soft” bipolarity is defined, from 3%–6% of the population reports bipolar-type symptoms and falls into this category (21–23). Should all of these people be considered to have BD? Should all of these people take medications? Are some of these individuals representative of a group for whom complementary treatments could be critical and should be studied? We do not know the answer to these questions.

Final revisions of DSM-5 are still a work in progress; however, the Bipolar Disorder Sub-Workgroup in collaboration with the full Mood Disorder Workgroup for the DSM-5 has proposed to update BD and depressive disorder in a number of ways. This set of proposals is being evaluated by multiple committees, and many of the proposed changes are still undergoing feasibility testing in the field. What follows are the current proposals. Additional information about DSM-5 can be found at www.dsm5.org.

Duration of hypomania

One of the key issues discussed by the Mood Disorder Workgroup was whether to preserve the current duration criterion of 4 days for hypomania or to shorten the criterion in DSM-5 to 2 days. Extensive review of the literature and new evidence from a large multisite study supports the continuance at 4 days (22–24). The Mood Disorders Workgroup for DSM-5 is making the recommendation to maintain Criterion A for an episode of hypomania at 4 days' duration. The proposal is as follows:

A distinct period of abnormally and persistently elevated, expansive, or irritable mood and abnormally and persistently increased activity or energy, lasting at least 4 days and present most of the day, nearly every day, that is clearly different from the usual non-depressed mood (24).

Extensive consideration was given to decreasing the number of days to define a hypomanic episode. If the duration were decreased for the defining of an episode of hypomania to fewer days, it was estimated that the “prevalence” would increase by 2–3 times (26). There was a lack of compelling data to change the duration criteria (27); in fact, many large studies examine subthreshold bipolarity without consideration of duration (28–30). External validators used to support the specificity of a diagnosis or disorder support maintaining the hypomania duration criterion at 4 days (22, 23). The recent “BRIDGE” study supporting this criterion is a large international study of patients currently experiencing a major depression episode for which an evaluation was made of various cut points for duration of hypomania between 1, 2–3, and 4–6 days. External validators, including first-degree relatives with a mood disorder, early adult onset, recurrent mood episodes, mood lability, seasonality, and history of suicide attempts support greater specificity of BDII by maintaining a 4-day duration for a hypomanic episode (22, 23). There was a striking linear relationship between many of the external validators evaluated and the duration of the hypomanic symptoms. For example, the percentage of the population reporting a family history of mood disorders increased as number of days of hypomania increased. The difference between requiring 2–3 days versus 4–6 days in these validators was notable, in many cases improving by 20% or more.

Although BDII is well-recognized and studied in North America, this diagnosis is still classified and recognized by the ICD-10 (31) as “Bipolar Other.” Shortening the criterion duration of hypomania could decrease the specificity of the diagnosis and add to the periodic skepticism regarding the integrity and validity of bipolar II as a separate diagnostic category (6, 12, 20).

Mixed features specifier

Perhaps one of the most significant proposals for DSM-5 is the change under consideration for mixed mania or mixed episodes in the mood disorders. Currently, mixed episode, often referred to as mixed mania, is defined as an individual meeting the full criteria for both mania and depression in the same time period and experiencing symptoms relatively simultaneously (3).

Since DSM-IV stated these criteria in 1994, the relative rarity of individuals meeting the full criteria juxtaposed to the high frequency in BD of mania with some degree of depression symptoms or depression with hypomanic symptoms has become clear and alternative definitions have been proposed (28, 32–40). In particular, the occurrence of mania, hypomania, or depression with subthreshold nonoverlapping symptoms of the opposite pole is at least as common as that of the “pure” classic euphoric mania or depression; furthermore, awareness of mixed symptoms may be important for treatment choices and carry potential implications for the course of illness (28, 41–43). The change proposed in DSM-5 is to capture mixed symptoms as a mixed features specifier (25) (Table 1).

Importantly, this specifier is proposed to apply not only to patients with BD but also to patients with major depressive disorder (MDD). This change is proposed on the basis of recent studies, showing by conservative definitions, that 20% of patients experiencing a depressed episode experience subthreshold hypomanic symptoms (28–30). A recent study of the National Institute of Mental Health (NIMH) Depression Collaborative cohort suggests that such symptoms may in some cases be predictive of a switch to BD (28–30). If more liberal definitions of subthreshold hypomania are used, such as requiring fewer numbers of symptoms, then the incidence goes higher and closer to 40% of major depressive episodes have some degree of bipolar hypomanic symptoms (21–23, 28).

It is also important to note that the proposed symptoms from the opposite pole are nonoverlapping. Symptoms for which it is more difficult to discern causality and define whether they are related to depression or mania, such as irritability or distractibility, were not included.

If this current proposal is accepted and becomes part of DSM-5, it will be the first time the bridge between BD and MDD will be formally recognized. Some studies suggest that individuals with MDD with mixed features may be individuals at greater risk for developing BD in the next decade (28–30). Little well-controlled research on treatment of patients with MDD and mixed features has been completed, although naturalistic observations raise the issue whether antidepressants would be a first-line treatment or not. The answer can only be “maybe” before formal treatment trials are performed.

The Mood Disorders Workgroup for the DSM-5 is making the recommendation for the addition of “activity or energy” to Criterion A for hypomanic and manic episodes. The proposal for the updated criterion is as follows:

A distinct period of abnormally and persistently elevated, expansive, or irritable mood AND abnormally and persistently increased activity or energy. Symptom lists (Criterion B) would be essentially not changed (25).

With Criterion B, it is currently possible to meet DSM-IV-TR criteria for bipolar I disorder without a change in activity or energy. The addition of requiring increased activity/energy for Criterion A will make explicit the requirement that this hallmark symptom of mania and hypomania be present to make the diagnosis. Addition of this feature is supported by a number of studies suggesting that activity or energy is at least as important as mood, and some studies argue that activity and energy are more important than mood (2, 22, 23, 44–49).

Bipolar disorder not otherwise specified revised to include specific types

Still under discussion for DSM-5 is how all “not otherwise specified” categories will be captured and coded. The Mood Disorders Workgroup for the DSM-5 proposes to revise BD not otherwise specified (BD NOS) to allow subcategories to be coded and to be specific in definition. This is an important change from DSM-IV-TR, in which not otherwise specified was defined by “examples include.” This proposed change will allow the capture of the phenomenon of patients experiencing hypomanic symptoms with notable change from usual behavior that are less than 4 days in duration or of a lower symptom count but full duration than is required to meet the criteria for a full hypomanic episode. In contrast to the proposed mixed features specifier (25, 50), for which symptoms occur simultaneously, symptoms for BD NOS by definition occur at separate times (25, 50) (Table 2).

Major breakthroughs have occurred in the last 60 years (lithium), 30 years (anticonvulsants), 20 years (U.S. use of clozapine and early SGAs), and even 10 years (more SGAs and lamotrigine). The rise of SGAs and the attendant marketing, media, and medical attention brought a level of focus to BD that had never been seen. The traditional schizophrenia-based markets became overshadowed by the broadly defined BD market. First, the discovery that 50% or more of patients in a BDI manic episode, including those without psychotic features, responded to SGAs and without the side effects of the first-generation antipsychotics, paved the way to consider this group of medications for broader use (51–53). Although use of the first-generation antipsychotics to manage BD was common before the development of SGAs, the utility of these agents was limited by the long-term side effect profile and limited efficacy, if not worsening of depression (2). We observed a sea change when evidence of the efficacy of SGAs in depression was shown and our concept of the uses of SGAs in medication management expanded (54, 55).

The new and important concept of SGAs as both antimanic and in some cases antidepressant emerged apart from the concept of antipsychotics primarily treating psychotic symptoms. That SGAs have an impact on mood unrelated to the presence of psychotic features is notable. However, for BDI we have not satisfactorily solved the issues of depression or residual depression. The limit of our evidence base is quickly reached, and, in particular, the evidence rarely approaches the often complex treatment regimens seen clinically (56, 57). The impetus to extend these evidence bases is limited by funding constraints. Conservatively, about 2% of the U.S. population has BDI, BDII, or more severe bipolar spectrum illness. Yet we reach the limit of having multiple choices for evidence-based treatments almost as soon as we pass acute mania for BDI, and acute depression studies, although important and groundbreaking, focus on monotherapy and provide few options.

Much of this increased knowledge was driven by the development of medications for U.S. Food and Drug Administration approval and thus by pharmaceutical companies. Historically pharmaceutical companies shine in drug development, but the exploration of the full potential of a medication requires federal and other funding sources. When treatment studies are driven primarily by the needs of bringing a medication to market, it is unlikely that a full range of studies will ever be completed to explore the complex medication regimens seen in clinical practice. Recent efforts to explore more naturalistic treatments were supported by NIMH and the Stanley Medical Research Institute.

These extensive pharmaceutical development programs were juxtaposed to NIMH funding of large multisite clinical networks supporting studies addressing treatment of MDD, BD, and schizophrenia. Some key results have emerged from these studies. In the NIMH Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) studies, our understanding of the limited response to antidepressants was extended, awareness of co-occurring conditions in BD was enhanced, and appreciation of the overall symptoms and course of illness was increased (35, 58–61).

The Stanley Medical Research Institute funded the Stanley Foundation Bipolar Network, now known as the Bipolar Treatment Network, from whose work important understanding emerged about BD (62, 63). From this Foundation-funded, long-term study, key factors were observed, such as the association of early abuse with greater severity of illness, including an earlier onset, complex symptom presentation, and the striking co-occurrence of additional axis I disorders (32, 64, 65). In addition, some of the earliest recognition of the impact of subthreshold depression symptoms and both the presence and impact of subthreshold manic symptoms during a depressive episode were noted (66, 67).

Treatment studies have slowed with the decrease of new drug development, leaving the study of better and varied uses of current medication yet to be done and key questions yet to be answered. For BDII, some of the most fundamental questions are as yet unaddressed. Should all patients with BDII receive medication? How does lithium compare with an antidepressant for acute depression? What is an appropriate long-term treatment program for BDII (66, 67)? There have been a number of studies by Amsterdam and colleagues (69, 70), supporting the theory that for at least some individuals with BDII an antidepressant may be the correct medication choice. Likewise, there are literature findings supporting lithium and more recently quetiapine as key treatment options (2, 6, 71).

The critical clinical questions relating to BDI and BDII are extensive, and although some evidence exists for the potential needs of each disorder, limited evidence exists to address the full range of patient needs. It is critical to maintain the awareness that these are severe and lifelong illnesses associated with significant suffering and personal, family, and societal impact. It is important to correctly diagnose and not to overdiagnose because of a few symptoms or to underdiagnose because of a lack of recognition of hypomanic or manic periods. Further studies are needed in medication and other therapies addressing patients with BDI and BDII.

The questions for BD NOS, arguably accounting for the largest group of people with potentially severe or at risk for developing more severe illnesses, are yet to be posed (21–23). Empiric and prospective treatment studies are needed to add to the already available literature to more formally define and explore treatment options for patients with BD NOS.

There is a current push toward personalized medicine, arguably a key and important direction for the next leap in treatment both for medicine and psychiatry in general. In some studies and new experimental models, work is being done to further this emerging field (72, 73). There is optimism that newer tools including whole genome studies, protein and mRNA expression, epigenetics, stem cell work, and “personalized” tissue culture systems will provide the avenue to not only increase our understanding at a cellular level of disease mechanisms but also provide individualized medicine.

These efforts are, without doubt, central and part of the revolution we are currently experiencing in medicine. However, in the process of the leap to genes and basic cellular components we have left behind the patients currently dealing with these illnesses and an inadequate evidence base to guide treatment. It is clear that both approaches are needed. Federally supported clinical trials are still needed, just as the scientific agenda is advanced addressing underlying mechanisms and developing personalized medicine options. Are there clues indicating other avenues that could substantially affect BD and MDD? Rather than just medications, are there other modifying biological factors that could have an impact on the course of illness? The article in this issue of FOCUS by Dr. Yang reviews recent studies providing clues into the biology of BD.

Possible role of inflammatory and cellular processes in bipolar disorder

One of the emerging areas of medicine is the putative role of inflammation throughout the body, including inflammation in the development of cardiovascular disease and inflammatory causes of narcolepsy (74–76). Results from The Netherlands of a putative mRNA inflammatory signature observed to be significantly different in healthy control subjects versus patients with BD suggest one future direction deserving of further explanation of bipolar disorder pathophysiology (77, 78). Just as secondary messenger systems have been implicated as having a role in the system of BD including destabilization—an imbalance corrected in some with lithium—potential mitochondrial dysfunction, a cellular dysfunction, may play a key role in BD (79–81). It is to be hoped that as our understanding of these important fields increases, it will open the way to novel treatments. Assessing this type of issue may, in fact, be the ultimate personalized medicine because each individual may have a unique, detailed inflammatory or cellular dysfunction profile.

An overview of changes proposed for DSM-5 and the state of treatment studies for BD was provided. The newer fields not touched on in this update are significant. We can anticipate revolution in the role of epigenetics in our thinking regarding the biology of BDs. Findings in this field are likely to be very much a part of the revolution of not just psychiatry. On the other side of this explosion of study, there are targeted interventions yet to be developed that may modify the expression of genes, profoundly changing the rate of development and course of BD.

Over the past 40 years significant progress has been made in our knowledge and understanding of this complex disorder. Valuable treatment interventions are now available, and many of these treatment options have been developed in the last 20 years. However, a gap exists in our understanding of the biological causes and the expression of the cluster of symptoms termed BD. Scientific efforts must continue in biological aspects and new treatment development to more fully understand this disorder and to adequately treat patients.

I thank Iola Gwizdowski for her support and editing in the preparation of the final manuscript.