Bipolar disorder suggests two discrete states, depression and mania. Yet episodes combining depressive and manic elements are not uncommon. Understanding these mixed states will improve our ability to define what constitutes an episode of bipolar disorder and the relationship between episode and illness course characteristics.
The first challenge in the investigation of mixed states is to define them. It is easy to regard depressive and manic states as opposites. Yet it has long been known that manic and depressive symptoms are not mutually exclusive (1). Manic and depressive symptoms are not generally negatively correlated—whether subjects in a study are mainly depressed (2), manic (3, 4), or either (5, 6).
Combinatorial models for mixed states posit essentially linear combinations of depressive and manic features. Emil Kraepelin (1) proposed combinations of elements that can each be considered depressed or manic. A second noncombinatorial model, based on proposals by the Vienna school (7), emphasized the underlying mechanisms of mixed states.
Mixed states significantly affect clinical course and treatment. Understanding them will enhance our grasp of bipolar and related disorders. Here, we concentrate on operationalizable definitions of mixed states, addressing the development of the concept, the symptomatic structure, the relationships between susceptibility to mixed states and course of illness, and the practical criteria for mixed states based on these factors.
The idea of manic-depressive illness, including mixed states, originated with classical physicians, especially Hippocrates (ca. 460–370 BCE) and Aretaeus of Cappadocia (ca. 30–130 CE) (8). Perhaps the first modern psychiatrist to describe mixed states systematically was Johann Christian August Heinroth (1773–1843), who categorized mental disorders as exaltations (“hyperthymias”), depressions (“asthenias”), and mixed states of exaltation and weakness (“hypo-asthenias”), further divided into mixed mood, mental, and volition disorders (9).
Emil Kraepelin (1856–1926) systematized previous observations, positing three elements of behavior that could be increased (manic) or decreased (depressed) independent of the others: mood, thought, and volition. Mixed states (“Mischzustande”) or mixed forms (“Mischformen”) were the six combinations of depressed and manic elements: depression with flight of ideas (“ideenfluchtige Depression”), excited depression (“erregte Depression”), depressive-anxious mania (“depressive oder angstliche Manie”), unproductive mania with poverty of thought (“ideenarme Manie”), inhibited mania (“gehemmte Manie”), and manic stupor (“manischer Stupor”). He proposed that mixed states were driven by a mechanism resembling hyperarousal and represented a more severe form of bipolar disorder. Wilhelm Weygandt (1870–1939), a pupil and colleague of Kraepelin, published the first book on mixed states, On the Mixed States of Manic-Depressive Insanity (10), in which Weygandt introduced the term “agitated depression” (“agitierte Depression”).
Kraepelin further classified mixed states temporally: either transitional between depression and mania or autonomic. Autonomic forms were the most unfavorable manifestations of manic-depressive insanity, with longer course, more chronicity, and longer episodes (1, 10)—confirmed 100 years later, as discussed below.
Kraepelin’s basic elements were derived from observation alone without access to the neurobiological or clinical information now available. Also, he proposed no method for measuring whether an element was “depressed” or “manic.” Therefore, we may accept Kraepelin’s basic approach, but we must look critically at how to apply it.
In Clinical Psychopathology (11), Schneider wrote, “we no longer believe in mixed states.…[W]hat might give the appearance of this is really an actual change or reversal, insofar as we can call the situation cyclothymic at all.” Berner et al. (7), however, developed the Vienna research criteria based on the “dynamic” concept proposed by Janzarik (12), which involved biological mechanisms governing drive and emotions. The Vienna criteria delineate a mixed affective subtype with sustained instability, the “persistent presence of a drive state contradictory to the mood state and/or the emotional resonance” (13, p. 164).
DSM-III, DSM-IV, and ICD-10 adopted a pragmatic, descriptive approach that emphasized manic episodes with prominent depressive features. In the past 30 years, a rebirth of interest in mixed states has led to the investigation of mixed mania (mania plus less than syndromal depression) and, more recently, mixed depression (major depression plus less than syndromal mania). This approach addressed depressive and manic episodes separately, regarding mixed states as subtypes of these episodes rather than in terms of episode components (like Kraepelin) or underlying mechanisms (like Berner). One current challenge is to develop models in which reliably observable clinical characteristics can drive research on the mechanisms and treatments of these states.
Centuries ago, clinicians realized that manic and depressive characteristics could combine in the same episode, potentially representing unusually severe illness, and that these combinations appeared to result from some driving force. More recently, mixed states were conceptualized as forms of depression or of mania. In the next section, we explore carefully observed clinical properties of mixed states, reviewing characteristics that differentiate mixed from nonmixed episodes and comparing mixed states derived from manic or depressive episodes.
Kraepelin formulated mixed states as freestanding or transitional episodes rather than as depressive or manic subtypes (1). However, mixed states have generally been studied as subtypes of manic or depressive episodes. Below, we review the properties of predominantly manic or depressive mixed states and their potential convergence.
While current definitions require full syndromal mania and depression, episodes with subsyndromal depression differ in treatment response (14, 15), and mixed hypomanic episodes appear common (16). Definitions of mixed mania range from narrow definitions in DSM-IV or ICD-9 to definitions based on subsyndromal depressive symptoms (including the Cincinnati criteria of at least three depressive symptoms in mania ); treatment response differs with two depressive symptoms (14).
Although they were not addressed in DSM-IV or ICD-9, depressive episodes with concurrent hypomania or mania appear common (18–20). The range of definitions for depressive mixed states resembles that for mania. Clinical studies suggest that features distinguishing mixed depressive episodes emerge at low levels of manic symptoms (2) and are often missed (21).
Some patients may experience hypomanic or manic symptoms only during depressive episodes (19, 22), especially early in the course of illness (20). These patients had early onset, frequent episodes, and family history of bipolar disorder, resembling patients with freestanding hypomania (18). Manic symptoms in depressive episodes have been shown to be related to severe illness course; receiver operating characteristic analysis revealed a threshold at modest manic symptom levels (Mania Rating Scale score, 6) (2). Large prospective studies have demonstrated that subsyndromal manic symptoms during depressive episodes in patients diagnosed with major depressive disorder are associated with family history and illness course characteristics resembling bipolar disorder (20) and with eventual diagnostic change to bipolar disorder (23).
Independent of predominant polarity.
Kraepelin’s classification of mixed states was not specific to depressive or manic episodes (1). Yet, mixed states in the context of affective episodes in general have rarely been addressed. The data we review below suggest that regardless of dominant polarity, characteristics of mixed states emerge with two or three symptoms of opposite polarity, although this has rarely been studied directly.
Manic episodes with two or more nonoverlapping depressive symptoms were found to have a poorer response to lithium than those with fewer depressive symptoms (14). Greater anxiety emerged with two symptoms; suicidal behavior and early onset of illness emerged with three or more (6).
The presence of two or three manic symptoms (defined to not overlap with potential depressive symptoms) in a major depressive episode confers properties generally associated with bipolar disorder (history of freestanding manic or hypomanic episodes and family history of bipolar disorder) and mixed states (severe course and co-occurring disorders) (19, 20). Two symptoms may be too sensitive, and three symptoms may be too specific (19).
Affective symptoms in mixed states.
Manic symptoms that characterize mixed manic episodes include greater mood lability and irritability and decreased grandiosity, euphoria, pressured speech, and need for sleep compared with nonmixed episodes (24). Manic symptoms have been reported as less severe (25), more severe (15, 26), or comparable in severity (27) in mixed compared with nonmixed manic episodes.
Depressive symptoms shown to be associated with mixed manic episodes included dysphoric mood, anxiety, excessive guilt, and suicidality (24). Mixed manic episodes have been found to be similar to agitated depressive episodes without other manic symptoms in severity of depression and anxiety, with more severe agitation, irritability, and cognitive impairment (28).
Manic symptoms that have been identified in mixed depressive episodes include irritable mood, distractibility, racing thoughts, and increased talking (19). In the Systematic Treatment Enhancement Program–Bipolar Disorder study (N=1,380), distractibility, racing thoughts/flight of ideas, and agitation were the most common manic symptoms during depressive episodes (29). In patients originally diagnosed with major depressive disorder, decreased need for sleep, increased energy, and increased goal-directed activity were found to predict a diagnostic change to bipolar disorder (23).
Depressive symptoms in mixed depressive episodes.
Unlike the case for manic episodes, we know of no reports on depressive symptoms distinguishing mixed from nonmixed depressive episodes. The severity of depressive symptoms appears similar in depressive episodes with or without concurrent manic symptoms (2, 20).
Symptoms regardless of polarity.
Greater anxiety was found to be predicted by at least two depressive symptoms in mania or at least one manic symptom in depression. Discriminant function analysis identified worry, negative self-evaluation, increased energy, visible hyperactivity, and racing thoughts as associated with mixed states, regardless of polarity (6).
Nonmood Symptoms in Mixed Episodes
Mood is only part of depressive and manic states. Nonmood symptoms, including anxiety, agitation, or psychosis, could be basic to mixed states. Specificity of relationships to mixed states may differ between predominantly manic or depressive mixed states because their roles in depressive and manic syndromes differ.
Kraepelin used “depressive” and “anxious” mania interchangeably, describing the patients as “frantically anxious,” reflecting general hyperarousal (1). Subsequent evidence confirmed this. In mania, high anxiety scores are associated with depression (30) and lithium resistance (15). Anxiety scores were found to be correlated with depression scores in mania (6) and were minimal in nonmixed mania (6, 15, 28). Anxiety appears to be a core symptom of mixed manic episodes (31).
Severity of anxiety in manic mixed states resembles nonmanic agitated depression (28), associated with increased tension and intensity of affect (32).
Unlike mania, in which anxiety is basically limited to mixed episodes, anxiety is pervasive in depression, mixed or not (28). Sato et al. (33) wrote that anxiety is “unlikely to play a major role in the core phenomenological features of mixed depression.” Yet, in bipolar depressive episodes, the severity of anxiety correlates with mania (6) as well as depression scores (2). Therefore, anxiety is strongly related to depressive mixed states, although its diagnostic utility is limited because anxiety is also prominent in nonmixed depressive states.
Anxiety has been shown to correlate with depressive symptoms in manic episodes, with manic symptoms in depressive episodes, and with the degree to which symptoms were mixed regardless of polarity (6). This pervasive role of anxiety is consistent with mixed states being driven by hyperarousal (1, 34).
Agitation, prominent in depressive, manic, and mixed states, includes two basic disturbances (35). The first is painful inner tension, which produces increased non-goal-directed motor activity (pacing and wringing of hands) and is often prominent in depressive episodes, regardless of whether manic symptoms are present (28, 35). The second is increased, poorly regulated goal-directed activity, often with irritability and impatience, which is prominent in manic episodes, regardless of whether depressive symptoms are present (28, 35). Mixed episodes, whether predominantly depressive or manic, generally include both aspects of agitation (1, 28).
It has been proposed that agitated depressed states are, of necessity, mixed (36), but this may not be the case. The Zurich Study, following 591 individuals over 20 years, found both agitated and retarded depressions associated with bipolar disorder and anxiety. Combined agitated and retarded depressions occurred in bipolar disorder, while purely agitated depressions characterized unipolar disorder (37). These findings, along with other research (28, 35), show that some but not all agitated depressive states are bipolar (37).
Manic episodes have a high prevalence of psychotic symptoms, so severity of psychosis is not generally greater in mixed than in nonmixed mania (15, 38). Multivariate statistics identify a psychotic manic subtype that is distinct from depressive mania (3, 27, 38, 39).
As with mania, factor analysis of bipolar depressive episodes yields a psychosis/delusions factor (33). The prevalence of psychotic symptoms is lower in depression than in mania, so psychotic symptoms correlate with manic symptom severity in bipolar depressive episodes (2) as well as with the degree to which affective symptoms are mixed (6). Data show that psychotic features are more likely to be mood incongruent in mixed than in nonmixed bipolar depressions (40).
Structure of Episodes: Multivariate Analyses
We reviewed studies that used multivariate analyses of symptoms in manic episodes, depressive episodes, and across both manic and depressive episodes, focusing on prospective studies with validated symptom or behavioral measures.
Manic and mixed manic episodes.
Based on factor analysis of manic episodes, cluster analysis characterizes subgroups, including mixed states, as summarized in Table 1. The results are basically consistent across analyses. The factors included depressive appearance and behavior, manic appearance and behavior, irritability, and psychosis. The depression factor scores were sometimes biphasic (3, 27); other factors were normally distributed. Cluster analyses consistently revealed “classic” euphoric, delusional/psychotic, hostile/irritable/dysphoric, and Kraepelinian depressive/anxious mania subtypes.
TABLE 1.Symptomatic Structure of Manic and Mixed-Manic Episodesa
| Add to My POL
|Cassidy et al. (3, 24)||Self-rating derived by authors from SCID, ROC, GOM||Depressed mood,b psychomotor acceleration, psychosis, increased hedonia, irritable-aggression||ROC: mixed > one depressed mood symptoms; GOM: hypomania, mania, delusional mania, two mixed states|
|Dilsaver et al. (4)||SADS||Manic activation, depression, sleep disturbance, irritability-paranoia||Pure; two mixed states: depressive, irritable-dysphoric|
|Sato et al. (39)||37 symptoms from SADS||Depressed mood, irritable aggression, insomnia, depressive inhibition, mania, lability aggression, psychosis||Pure, aggressive, psychotic, depressive|
|Swann et al. (27)||SADS||Impulsivity, hyperactivity, anxious pessimism,b distressed appearance, hostility, psychosis||Classic, delusional, depressive, irritable dysphoric|
|Harvey et al. (95)||HAM-D, SADS, MRS||Five-factor solution in pretreatment subjects|
Depressive and mixed depressive episodes.
Table 2 summarizes factor analyses of depressive episodes. Here, we found fewer cluster analyses of the type in Table 1. These studies focused on the relationships between severity or number of manic symptoms and clinical characteristics. Activated (41) or emotionally hyperreactive (32) depressed patients potentially represent mixed depressive states. Measuring these characteristics may improve detection and monitoring of mixed states.
TABLE 2.Symptomatic Structure of Depressive and Mixed-Depressive Episodesa
| Add to My POL
|Henry et al. (32)||Multidimensional Assessment of Thymic States; beyond conventional affective symptoms||Emotional reactivity, psychomotor speed, psychomotor function, motor activity, sensory perception||Two depressive types, one with emotional hyperreactivity|
|Benazzi (96)||SCID||Mania-related factors: irritability, increased mood, motor overactivity||Previous diagnosis: ≥2 manic symptoms, 73% bipolar, 42% unipolar; ≥3, 46% bipolar, 8% unipolar|
|Biondi et al. (41)||Author-derived scale, broad range of behavior beyond conventional mood symptoms; MMPI-2||Unipolar depressive patients: depression, anxiety, activation||Activation in subjects with diagnosis of unipolar depression suggests misdiagnosis|
|Sato et al. (33)||43 symptoms from SADS||Vegetative signs, depressive retardation/reduced feeling, hypomania, anxiety, psychosis, depression/hopelessness||Hypomanic factor in depression, ? of mixed depression or “latent hypomania”|
Episodes across polarities.
Multivariate analyses across polarities are summarized in Table 3. Depressive or manic symptom factors can exist independently or in combination (42, 43). Contrary to Kraepelin’s negative relationship between manic and depressive symptoms (1), these analyses, like those reported in Tables 1 and 2, imply that manic and depressive symptoms are orthogonal (5, 6). The results support Kraepelin’s dimensional model but suggest different dimensions. They also identify constructs, including dysphoria, that may be useful in identifying mixed states regardless of polarity (34). One study (6) developed a measure of the extent to which symptoms were mixed independent of dominant affect—the Mixed State Index, essentially the product of normalized depression and mania scores. A high score on the Mixed State Index was related to severe course of illness and suicidal behavior. Factor analyses of personality scales across states identified subjects at risk for mixed states, personality disorders, and co-occurring disorders (44).
TABLE 3.Studies Across Depressive and Manic Statesa
| Add to My POL
|Meyer and Hautzinger (42)||CES-D plus SCID mania items||2,059 healthy subjects, two factors: dysphoric-depression, euphoric-activation||Bipolar-like structure including “mixed” even in healthy subjects|
|Adler et al. (43)||Affective Self-Rating Scale: depressive, manic, mixed||Mania and depression factors||Mixed subgroup with both factors elevated|
|Bertschy et al. (34)||MINI plus dysphoria||Dysphoria: ≥3 of inner tension, irritability, aggressive behavior, hostility||Dysphoria in 17.5% of pure depression, 22.7% of pure mania, 73.3% of “full mixed”|
|Perugi et al. (44)||TEMPS-A (temperament scale) plus interpersonal sensitivity, separation anxiety||Range of patients with bipolar disorder; cyclothymic-sensitive versus hyperthymic factors||Cyclothymic-sensitive associated with comorbidities and mixed states|
|Swann et al. (6)||SADS||Measure of extent to which symptoms were mixed across polarities (MSI)||MSI correlated with anxiety and psychosis, higher with early onset and comorbidities|
During mixed states, depressive and manic symptoms cover the full range of severity, each potentially varying negatively, positively, or independently over time (26). The characteristics of depressive and manic mixed states are similar, and they emerge as the severity of the opposite polarity increases. This raises the question of whether a group of patients with bipolar disorder is susceptible to mixed states. This susceptibility may be reflected in the course of illness.
The relationships between mixed features and course of illness have been studied through recurrence-related characteristics, such as age at onset, stability of mixed states, and number of episodes, and complications, such as co-occurring disorders, suicide risk, and treatment resistance.
In mixed mania, the results from studies of age at onset are inconsistent (17, 45). In general, subjects were those in whom any manic episode was mixed, but one negative study focused on mixed first episodes (46). Onset was earlier in individuals with at least one mixed episode over a 10-year period (47). Interpretation is limited by variable definitions of age at onset and of mixed states. Specificity of early age at onset for mixed states compared with stronger genetic liability to bipolarity is unclear. In a small study of age- and gender-matched bipolar patients, after controlling for age at onset, mixed states occurred only in those with a family history of bipolar disorder (48).
Studies of mixed depressive patients found consistently earlier age at onset than in nonmixed depressions (2, 29, 49). Mixed features across polarities were also related to early onset, independent of episode type (6). This may reflect increased severity of early-onset bipolar disorder in general (50).
Lifetime Number and Frequency of Episodes
Patients with mixed manic or depressive states have more lifetime episodes of illness than those without mixed states (27, 47, 51–54). Mixed states can also present as fewer episodes of longer duration (52). In either case, individuals with mixed states experience longer periods of mood instability.
Time to syndrome resolution is longer for mixed episodes than for depressive or manic episodes (49, 52, 55, 56). One study of 143 mixed and 118 manic episodes found that mixed episodes were longer, with lower rates of interepisode remission (52). After remission, mixed states relapsed sooner (57). Accordingly, course and prognosis of mixed mania are worse than for nonmixed mania (17, 45).
Stability of Mixed States
The possibility of a subtype of bipolar disorder characterized by susceptibility to mixed states can be addressed by prospectively investigating outcome after an index mixed episode or by retrospectively examining whether a current mixed state is associated with a higher than expected frequency of past mixed episodes. This susceptibility could be present from the outset or could emerge over the course of illness.
In 247 manic patients (97 mixed and 150 nonmixed) followed for 24 months after hospitalization, index mixed states were followed by 12 times more subsequent mixed states, 6.5 times more depressive episodes, and 69% more dysthymia than nonmixed manic index episodes. In contrast, nonmixed manic index episodes were followed by 10 times more manic episodes, six times more hypomanic episodes, and 35% more psychotic illness than were mixed index episodes (58). In 68 prospectively evaluated subjects with two discrete manic or mixed episodes, recurrence was true to type whether using a more (DSM-III-R) or less (Duke receiver operating characteristic criteria) inclusive definition for mixed mania (59). In bipolar depression, mixed depressive states were found to be relatively, but not absolutely, stable over 2 years (60). These studies support stability of mixed and nonmixed bipolar disorder, at least over 12–24 months.
Retrospective studies using DSM-III, DSM-III-R, the Pisa criteria, and the Cincinnati criteria for mixed states consistently found more histories of mixed episodes in patients with a current mixed presentation (52, 61, 62). Patients whose manic episodes were exclusively mixed had more previous mixed episodes than those with both mixed and nonmixed manic episodes (54).
Prospective studies, with the advantage of directly observed episodes, and retrospective studies, with the advantage of longer periods of observation, concur that susceptibility to mixed states is stable within individuals, for mania or depression.
Mixed mania combines depression with the impulsivity and hyperactivity of mania. Greater suicidal ideation, comparable to depressive episodes, was found in mixed relative to nonmixed mania (24, 39, 51, 62–65). A chart review of 184 inpatients with bipolar I disorder found higher past suicidality in patients with mixed manic episodes, defined by at least three depressive symptoms (65). Prospectively studied patients in mixed episodes had a higher rate of suicide attempts than those without mixed states, but not after adjusting for age at onset (47). Patients with exclusively mixed manic states had higher risk for suicide than patients whose history included nonmixed manic episodes (54). Suicidal ideation was found to be greater with at least three depressive symptoms in manic episodes not meeting DSM mixed-state criteria (66), and it correlated with depressive symptoms during manic episodes (64).
Mixed depressive episodes combine the hopelessness of depression with the impulsivity and hyperactivity of mania (67). Depressed bipolar patients with subsyndromal or presumptive DSM-IV-defined mixed episodes were found to be more likely to have attempted suicide than those without concomitant manic features (29). Greater anxiety in mixed states may increase suicide risk further (68). In one study, suicidal behavior was increased with “subsyndromal hypomania” during depression compared with purely depressive episodes; age at onset did not differ (20).
Unlike manic episodes, risk of suicide is already increased in nonmixed depression. Presence of manic symptoms adds to risk. Noneuphoric hypomanic symptoms have been shown to be associated with increased suicidality in depressed patients who had never experienced freestanding hypomania or mania (22). In bipolar depression, receiver operating characteristic analysis has shown increased suicide attempt history with a Mania Rating Scale score of 8, parallel to increased impulsivity (2).
Mixed states—depressive or manic—combine hopelessness with impulsivity and activation, leading to high risk for suicidal behavior. It is useful to know whether an individual is susceptible to mixed states, as emergence of mixed features in an initially nonmixed episode may increase the risk for suicide. This could occur pharmacologically (69).
Individuals with mixed manic or depressive episodes are more likely to have experienced early-life stressors and to have stress-related disorders (70).
Co-occurring anxiety disorders are more common in patients with mixed than nonmixed manic (63) or depressive episodes (71). This was confirmed by two large longitudinal studies: the National Epidemiologic Survey on Alcohol and Related Conditions (N=43,093) (72) and the Zurich Study (37). Co-occurring anxiety disorders are associated with a poor course of illness and increased service utilization (73).
Alcohol and substance abuse may be more prevalent in mixed than nonmixed manic episodes (2, 55, 74) and are associated with poor outcome and less favorable response to lithium (55). Differences in lifetime alcohol abuse, however, were not always confirmed (62). History of substance use disorder predicted greater mood instability in bipolar depressive episodes (75), possibly contributing to increased risk for mixed features and suicidal behavior.
Manic or depressive syndromes can be related to many medical and substance-related disorders. Diagnostic criteria for mixed states (discussed below) define them as primary affective syndromes, as do the articles cited in this review. As with other mood states, clinicians should be alert to the possibility that even with clear previous primary episodes, any episode may be related to another medical condition requiring treatment. This is especially true in mixed states, which may be more likely than other episodes to be associated with substance-related, nonpsychiatric medical, or neurological disorders (55).
Naturalistic studies, prospective and retrospective, have demonstrated that mixed states are difficult to treat, suggesting that mixed episodes respond poorly to lithium and other monotherapies (14, 55, 76). Placebo-controlled data generally derive from subanalyses of randomized clinical trials, whose enrollment criteria may have excluded patients with co-occurring conditions and suicidality (typical of mixed states); these studies have supported divalproex, carbamazepine, and second-generation antipsychotic agents (77). One study found a reduction in lithium effects with two or more depressive symptoms, without effect on response to divalproex; lithium was highly effective in episodes without depressive symptoms (14).
Mixed depressive episodes pose their own challenges. Akiskal et al. (22) proposed that mixed depressive episodes should not be treated with antidepressants because of increased activation and risk for suicidality. Compared with nonmixed depression, mixed depression had longer duration, worse outcome, and poorer response to treatments (49). In a large chart review study, Koukopoulos et al. (69) reported poor response to antidepressants in agitated, largely mixed, depressed patients. A large prospective study confirmed the poor response to antidepressants (78). In summary, antidepressants without mood-stabilizing treatments have limited efficacy and the potential for harm in bipolar disorder, and properly defined mixed depressive episodes are part of bipolar disorder.
Patients with mixed states have a poorer long-term prophylactic response to treatment in general (47). One question, given the high risk for suicidal behavior in mixed states, is the role of prophylactic lithium in patients susceptible to mixed states. Notably, data show that lithium appears to reduce suicidal behavior even when it is not effective in preventing affective episodes (79).
Patients who are susceptible to mixed states differ from those who are not—they have more episodes, increased suicidality, higher rates of co-occurring conditions, and poorer response to treatments. Similarly, patients with a more unstable form of bipolar disorder, including susceptibility to mixed states, are likely to be resistant to lithium (76). Patients with mixed states are more likely than other patients to have future mixed states. These illness-course characteristics, like the symptomatic characteristics described earlier in this review, appear consistent across mixed depression and mania. They raise questions of common mechanisms underlying mixed episodes and severe course of illness and of how, since they have important potential clinical and treatment consequences, mixed states can be identified efficiently.
Potential Mechanisms of Mixed States
Both Kraepelin (1) and the Vienna school (7) proposed that mixed episodes resulted from a pathological drive state, possibly related to emotional lability (32) and activation (41). Of the six Kraepelinian mixed states, the most prevalent are the activated states: depressive/anxious mania, depression with flight of ideas, and activated depression. Consistent with clinical studies, catecholaminergic function in mixed states has been shown to be greater than in nonmixed manic episodes (53, 80). These results, over nearly 100 years, converge to suggest that hyperarousal underlies mixed episodes.
Mixed states occur in the context of early onset, frequent episodes or prolonged affective instability, and increased stress-related (70) and substance use (74, 75) disorders. Mechanisms of this relationship between susceptibility to episode-related behavioral instability and a long-term unfavorable illness course represent an important target in understanding mechanisms and treatment targets in bipolar disorder.
Combined depressive and manic features could result from any of the following:
1. Superimposed physiological states: Mixed manic states combine the hypothalamic-pituitary-adrenocortical (HPA) activation of depression with increased catecholamine function exceeding nonmixed mania (80). Mixed mania is associated with increased HPA axis function, correlating with depressive symptoms. Low thyroid function has been associated with affective instability, but relationships between mixed states and thyroid function are inconsistent (81, 82). This implies a combinatorial structure, possibly driven by poorly regulated arousal and mood lability (32, 41).
2. Rapid alternation between mood states, such as ultradian cycling or extreme mood instability (83): Affective instability, including ultradian and other extremely rapid cycles, should be distinguished from episodes in which they occur, which can be prolonged (84). Similarly, mixed states should be differentiated from rapid cycling, which consists of frequent episodes of illness that could be manic, depressive, or mixed. Evidence for validators of rapid cycling other than frequent episodes has been inconsistent; most characteristics of rapid cycling appear as functions of episode frequency (85).
3. Transitions to and from mixed states: Such transitions do occur, although mixed states may also be stable and prolonged (52). Kraepelin suggested that some mixed states were transitions between depressive and manic episodes (1), but this may not generally be the case (86). Transitions between manic or depressive states and mixed states may be more common, as demonstrated in a longitudinal study (26). Polyphasic episodes, in which state changes occur without a stable period, have unfavorable illness course characteristics resembling mixed states (87).
In summary, mixed episodes are characterized by combined or rapidly alternating behavioral states, possibly driven by hyperarousal and increased catecholaminergic function. They are associated with early onset, recurrent course (or prolonged mood instability), early stressors, and substance use disorders. These characteristics suggest that mixed states are an expression of behavioral sensitization in bipolar disorder. Exaggerated behavioral sensitization could result from interacting genetic susceptibilities and environmental or pharmacological sensitizing stimuli.
Assessment of Mixed States
Need for comprehensive assessment.
Mixed states require clinical vigilance. Mixed features can be present in, or emerge during, any episode. Therefore, it is vital to monitor both depressive and manic features in any bipolar disorder episode, as well as in patient self-monitoring by mood charting. Distinct clinical features emerge with mild symptoms of the nondominant polarity, easily missed if not sought specifically. Identification of opposite polarity symptoms must account for symptoms that could overlap between depression and mania, such as irritability, agitation, anxiety, or insomnia. This is addressed in the Cincinnati criteria (17) and the Schedule for Affective Disorders and Schizophrenia–Change Version (SADS–C) (88). Comprehensive scales include the SADS-C, the Internal State Scale (89), and the Multidimensional Assessment of Thymic States (32). In addition to conventional psychiatric symptoms, emotional intensity and affective instability may contribute to the assessment and understanding of mixed states (90).
Depressive and manic symptoms may have different diurnal rhythms that could cause mixed features to be missed with once-daily monitoring (91). Monitoring should reflect maximum symptoms over a 24-hour period.
Potential ascertainment bias.
Detection and description of mixed states may be distorted by ascertainment biases, such as Berkson’s bias (92), where relationships between clinical characteristics may appear exaggerated if derived from a population with independently increased susceptibilities to them. Associations between manic and depressive symptoms would be exaggerated if individuals with both were artificially more likely to be included in a study. However, bipolar disorder, by definition, includes both depressive and manic states, and mixed affective symptoms have been demonstrated in nonclinical populations (Table 3) (42).
For example, manic and depressive symptoms have been shown to contribute independently to service use, implying that these symptoms are more independent than clinical studies suggest (93). Symptoms were measured at times remote to service use, and depressive and manic symptoms were not necessarily concomitant. Contributions of depression and mania may appear independent because they increase service use through different mechanisms. Interestingly, combined symptoms had larger effects than either alone.
Berkson’s bias would not affect the existence or structure of mixed states. It may contribute to apparent prevalence, which is not our focus. If mixed states occur with early onset and frequent episodes, Berkson’s bias could affect apparent rates of co-occurring substance use or anxiety disorders, which have similar illness course relationships. Berkson’s bias can be minimized with proper controls (94), but the potential should be taken into account.
Practical Identification of Mixed States
Tables 1, 2, and 3 suggest that symptomatic and illness course characteristics of mixed states emerge, in depressive or manic episodes, with two or three independent opposite polarity symptoms. The properties of predominantly depressive and manic mixed states converge. These characteristics can yield parsimonious criteria for mixed episodes, summarized in Table 4. The research criteria may be more detailed, including symptom scales, than optimal for general clinical use. Table 4 also includes practical clinical criteria. As with the specifier proposed by the DSM-5 Mood Disorders Work Group, the focus is on symptoms of the opposite pole that are unlikely to overlap.
TABLE 4.Proposed Research and Clinical Criteria for Mixed Statesa
| Add to My POL
|Manic or hypomanic, research||At least three nonoverlapping depressive symptoms from DSM-IV major depressive episode criteria, OR||Based on criteria of McElroy et al. (17) and treatment response data (14)|
|Score ≥10 on HAM-D (or equivalent on Inventory of Depressive Symptoms) modified to exclude agitation, irritability, and insomnia||Based on Stanley Foundation data (16) and factor/cluster analysis (27)|
|Duration of depressive symptoms: 1 week if mania and 4 days if hypomania||Current duration of DSM-IV mixed episode for mania|
|Manic or hypomanic, clinical||≥3 nonoverlapping depressive symptoms nearly every day during the episode (4 days for hypomania, 7 days for mania)||Depressive symptoms exclude agitation, insomnia, irritability, and distractibility|
|Major depressive, research||SADS mania factor score ≥8 (2)||YMRS numerically similar but SADS is designed to measure mania, depression, anxiety, and psychosis simultaneously|
|≥3 mania symptoms non-overlapping with depressive symptoms (19)||Excluding insomnia other than decreased need for sleep, agitation other than increased goal-directed activity, and irritability|
|Duration of manic symptoms: 4 days||Criterion for DSM-IV hypomanic episode|
|Major depressive, clinical||≥3 nonoverlapping manic symptoms nearly every day for at least 7 days of the episode||Manic symptoms excluding agitation, insomnia, irritability, and distractibility|