Bipolar Disorder: Defining Remission and Selecting Treatment

Bipolar disorder (BD) is a highly prevalent and complex medical syndrome of multifactorial origin. It is estimated that about 2% to 4% of the general population is affected by a bipolar spectrum condition. Among clinical samples of depressed patients, the estimated percentage of persons who screen positive for BD is about 10% to 35%.^1-3 The longitudinal course of BD is characterized by a low rate of recovery, a high rate of recurrence, and poor interepisodic functioning.⁴ The standardized all-cause mortality ratio among patients with BD is increased approximately 2-fold.⁵ Actuarial estimates currently regard BD as possibly the most costly category of mental disorders in the United States.⁶

The foregoing clinical portrait of BD has only an evanescent similarity to the notions of vecordia and manic-depression, which were articulated by Kahlbaum and Kraepelin, respectively. These and other investigators diagnosed and categorized cyclical mood disorders (ie, BD) primarily on the basis of a favorable symptomatic and functional outcome and the absence of a "dementing" course.⁷

Results from several recently published longitudinal prospective studies have provided a more refined composite of BD. The symptomatic structure of BD is composed largely of subsyndrome depressive and anxious symptoms that rapidly shift in polarity and severity. For most persons affected, the symptoms of BD pursue an inexorably unremitting course that is accompanied by a progressive increase in vulnerability to recurrence, neurocognitive impairment, and psychosocial dysfunction.

Mortality studies have documented an increase in all-cause mortality in patients with BD. A newly established and rapidly growing database indicates that mortality due to chronic medical disorders (eg, cardiovascular disease) is the single largest cause of premature and excess deaths in BD.⁵ The standardized mortality ratio from suicide in BD is estimated to be approximately 18 to 25, further emphasizing the lethality of the disorder.

This description of BD points to the need to invoke a chronic disease management model (CDMM) when treating individuals with BD. A CDMM is encouraged for any syndrome, disorder, or disease characterized by multifaceted illness presentation, psychosocial burden, and chronic course. This model of health care delivery emphasizes a multidisciplinary systems-based approach with interventions at macro- (ie, health care system), meso- (ie, clinic structure) and micro- (ie, patient) levels.

Most clinicians are able to immediately intervene at the micro-level on a daily basis. Defining critical end points (ie, symptomatic remission) and using evidence-based guidelines to inform patients making treatment decisions are the guiding principles of a CDMM. Clearly, quantifying and objectifying patient outcomes with symptom measurement tools hitherto has not been common practice in BD management, let alone in the field of psychiatry.⁸

The encompassing aim in managing a patient with BD is to help him or her achieve wellness. Wellness has been defined and operationalized along 3 dimensions: symptoms, functioning, and pathophysiologic change.⁹ For many chronic medical disorders in which the pathophysiology has been elucidated (eg, coronary artery disease, diabetes mellitus), the availability of biomarkers has served multiple purposes, including quantification of illness severity, tracking of therapeutic progress, and determination of illness activity. For example, a glycosylated hemoglobin measurement exceeding 6.5% would indicate insufficient glycemic control over the past 3 months.

Since BD, like all psychiatric disorders, has an unknown pathoetiology, no valid and reliable biomarker has been identified as a surrogate for illness activity. In the interim, mental health care providers are encouraged to track patient progress toward illness by quantifying and objectifying "surface-based" phenomena along symptomatic and functional domains.

In this context, it is somewhat surprising that the use of measurement tools has not been common practice in BD. Measurement-based care has been proved to enhance remission rates, functional outcomes, and quality-of-life measures in real-world patients who are depressed and are receiving treatment with conventional antidepressant medications.¹⁰ As a starting point, critical end points in mood disorders need to be established and the most effective therapeutic avenues need to be adjudicated.

Several multinational expert guidelines on the management of depressive disorders emphasize remission--an outcome that transcends response--as an achievable and more clinically relevant symptomatic end point. Residual depressive symptoms and incomplete remission are associated with early relapse, shorter duration between depressive episodes, chronicity and poor prognosis of comorbid medical disorders, increased use of medical services, and sustained elevation of suicide risk, as well as psychosocial and functional deficits.^9,11

The most frequently reported symptomatic outcome measure in clinical trials of antidepressant medications has been the patient's response to treatment, arbitrarily defined as a reduction of 50% or greater from pretreatment in total symptom severity. A categoric response to therapy that fails to achieve a fully asymptomatic remitted state constitutes an unsatisfactory outcome in that it includes patients with ongoing clinically significant disease activity.

Several definitions of remission in depression have been proposed; the most frequently cited definitions include a total Hamilton Depression Rating Scale 17-item (HAM-D-17) threshold score of 7 points or less, and a Montgomery Asberg Depression Rating Scale (MADRS) score of at least 10 to12 points. A limitation of this definition is the presumption that the HAM-D-17 threshold score or the MADRS score is--prima facie--evidence of remission of a depressive episode. Nevertheless, these defini-tions have provided a useful starting point for defining and operationaliz-ing the end of illness activity in depression and serve as guiding principles for researchers and clinicians treating BD.¹²

Defining critical end points in BD is daunting considering its pleomorphic and fluid clinical presentation, the multidimensionality of the syndrome, its persisting neurocognitive deficits in the absence of overt mood psycho- pathology, its comorbidity or co-occurring syndromes, and its functional impairment across multiple domains.

Enhanced self-reported function associated with illness activity (eg, hypomania) is unique to BD. The multidimensionality of BD presents a unique challenge insofar as a patient with BD who is depressed and who switches into hypomania with a conventional antidepressant medication would be considered in depressive remission but clearly is not in "bipolar remission." Conversely, a person with BD receiving conventional antipsychotic medication for manic symptomatology that subsequently switches into postmania depression is also not in "bipolar remission."

Study results now press the point that neurocognitive deficits are a primary feature of BD; they are highly prevalent and persist in the absence of overt symptomatology.¹³ Although disparate neurocognitive abnormali-ties have been reported, disturbances in attention, visual memory, and executive function are most consistently reported.

Neurocognitive deficits are of more than subsidiary importance. For example, their presence is associated with poor functional outcome, nonremission, and increased probability of recurrence.^14,15 It is conjectured that neurocognitive deficits may comprise an en- dophenotype in BD. If a patient with BD is no longer manifesting depressive or manic symptoms but has persisting neurocognitive deficits that detract from full functional recovery, would we conceptualize this person accurately as being in remission?

The topic of comorbidity in BD is increasingly urgent.¹⁶ The term "comorbidity" first appeared in the medical epidemiology literature in 1970, in reference to "any distinct additional entity that has existed or that may occur during the clinical course of a patient who has the index disease under study."¹⁷ Community- and clinic-based studies have documented the high lifetime prevalence of, and risks posed by, comorbidity in BD. For example, the National Comorbidity Survey reported that 95% of the respondents with BD met criteria for 3 or more additional lifetime psychiatric disorders.¹⁸ The Stanley Foundation Treatment Outcome Network reported that 65% of "patients with bipolar disorder also met DSM-IV criteria for at least one other comorbid lifetime Axis I disorder."¹⁹

Taken together, comorbid conditions among patients with BD are associated with several indications of illness severity and hazardous dysfunc- tion. For example, comorbidity is associated with more severe subtypes of bipolar illness, an earlier age at onset, mixed-state presentations, an intensification of symptoms, poor symptomatic and functional recovery, suicidality, diminished acute response to phar- macologic treatment, a decreased quality of life, and an unfavorable course and outcome.¹⁹

Medical comorbidity in the patient with BD has negative implications for remission, probability of recurrence, and illness trajectory. For example, migraine, cardiovascular, cerebrovascular, pulmonary, and infectious diseases are frequently cited comorbidities in the bipolar population (Figure).²⁰ Preliminary evidence further suggests that migraine disease may covary with bipolar II more often than with bipolar I, implying that the migraine-bipolar overlap may represent a subphenotype of BD. Compelling evidence now indicates that individuals with BD are differentially affected by obesity, metabolic syndrome, and type 2 diabetes.²¹ The presence of a metabolic abnormality in the patient with BD is associated with nonremission, increased proba- bility of recurrence, depressive episodes, suicidality, and poor functional outcome.²²

Mortality studies indicate that the standardized all-cause mortality ratio in patients with BD is about twice that seen in persons without BD.^5,23 The total number of excess deaths in BD is the highest for natural causes (eg, cardiovascular disease), underscoring the specific hazard posed by medical comorbidity. Taken together, the burden of illness attributable to both psychiatric and medical comorbidity in bipolar patients provides the driving force for conceptualizing comorbidity as a primary target for preventive strategies, remission definitions, and therapeutic interventions. Moreover, a synthesis of the medical and psychiatric comorbidity literature in BD allows for a testable hypothesis that co-occurring syndromes may pathophysiologically overlap with BD.

If BD has an overlapping biologic substrate with co-occurring syndromes, would it then be accurate to refer to a patient with BD as being in remission if that person exhibited minimal manic and depressive symptoms, yet was morbidly obese and had type 2 diabetes and hypertension? This scenario would be analogous to the patient with cardiovascular disease who is asymptomatic with normalized blood pressure, but continues to display abnormal lipid parameters and dysglycemia. Would this person be referred to as being in remission?

Researchers in BD have traded several phrases back and forth, including euthymia; remission; and syndromal, symptomatic, and functional recovery. A coherent synthesis and preliminary working definition for remission in mania has been proposed. Chengappa and colleagues²⁴ define remission as a Young Mania Rating Scale (YMRS) score of 7 or less at the end point, with no core item of the YMRS (ie, irritability, speech, content, and disruptive-aggressive behavior) having a score greater than 2. The remaining 7 items on the YMRS are to be scored 1 or less, and the total 21-item HAM-D score was to be 7 or less. As a proxy for global functioning, the Clinical Global Impression Bipolar Severity score should indicate that the patient is functionally recovered (ie, score of 2 or lower).²⁴

Although this definition of remission in BD is highly stringent and not consensually accepted, it provides a useful starting point for clinicians. For example, the primacy of an asymptomatic state in BD emanates from this definition. Psychosocial impairment in BD increases with each increment in depressive and manic symptoms. At a corresponding level of severity, depressive symptoms in the patient with BD are as impairing as (if not more impairing than) manic symptoms.⁴ In keeping with this view, it becomes axiomatic that the elimination of disease activity provides a patient with the greatest probability of functional recovery.

Taken together, these points establish a compelling argument for achieving remission; a prospect for future research is to validate definitions of remission with measurement tools that are likely to be accepted by clinicians working in real-world settings. In the meantime, clinicians should track patient progress and inform patients about treatment options at critical decision points that are ubiquitous in BD.

Over the past decade there have been several evidence-based, consensus-based, and combined evidence/consensus based guidelines for the treatment of BD. The expanding pharmacopoeia and the development of psychosocial treatment strategies for BD have provided clinicians with more opportunities to benefit individual patients. But how are clinicians expected to select and sequence treatment strategies in BD?

The use of treatment guidelines provides a vehicle to enhance the appropriateness, consistency, quality, and cost-effectiveness of the treatments provided by clinicians. In reality, most guidelines for BD are a hybrid of both expert opinion and consensus, which may be required because pivotal clinical trials in BD often include nonchronic, noncomorbid cases that produce results with minimal ecological validity. Moreover, most clinical studies in BD are pharmacologic monotherapy treatment trials that are primarily designed for product approval purposes rather than for addressing unmet clinical needs. An expert opinion is often required when determining the next course of treatment when index therapies are not sufficient.

A comprehensive review of each of the listed treatment guidelines is beyond the scope of this article. We recommend that clinicians familiarize themselves with at least one treatment guideline and incorporate the algorithm of that guideline into their clinical practice. Increasingly, guidelines are using the concepts of remission to help clinicians choose the next steps in the treatment approach.