. 5
( 68 .)


Table 1.4 Schizoaffective disorders (F25) according to Tenth Revision of the International
Classification of Diseases (ICD-10: World Health Organization, 1991)

G1. The disorder meets the criteria for one of the affective disorders (F30, F31, F32) of
moderate or severe degree, as specified for each category
G2. Symptoms from at least one of the groups listed below must be clearly present for most of
the time during a period of at least 2 weeks (these groups are almost the same as for
schizophrenia F20.0“F20.3)
G3. Criteria G1 and G2 above must be met within the same episode of the disorder, and
concurrently for at least part of the episode. Symptoms from both G1 and G2 must be
prominent in the clinical picture
G4. Most commonly used exclusion clause. The disorder is not attributable to organic mental
disorder (in the sense of F00“F09) or to psychoactive substance-related intoxication,
dependence, or withdrawal (F10“F19)

and permutations.™™ For clinical purposes, therefore, we use a bidimensional
categorical system for classifying the cross-sectional affective state of our patients
with bipolar disorder (McElroy and Weller, 1997). Patients can have various
combinations of various degrees (none, mild, moderate, severe) of manic and
depressive symptoms, thereby allowing for more accurate diagnosis and, hence,
more appropriate treatment.
20 A. Marneros and F. K. Goodwin

There are no systematic epidemiological studies on mixed states. The estimation
of their frequency is mainly based on studies of psychiatric inpatients and, to a
lesser extent, outpatients. Going back to Kraepelin, it has always been clear that
their frequency is dependent on the definition applied. Thus, Kraepelin (1899),
as well as Weygandt (1899), estimated that by applying broad definitions, their
frequency is very high: approximately 60%. Applying a narrow definition,
which requires the full symptomatology of melancholia and of mania, reduces
the frequency to about 20%. Thus, contemporary reviews of the prevalence rates
of mixed states in patients with bipolar disorder report a range between 5 and
70% (Goodwin and Jamison, 1990; McElroy et al., 2000). With a median of
about 43% (Goodwin and Jamison, 1990), exactly the same percentage was
found in the Cologne study (Marneros et al., 1991a, b). Further, the Cologne
study observed bipolar patients over 25 years and noted that only 1% of the
patients consistently had mixed states (Marneros et al., 1991a, b). It seems that
the frequency of mixed states is related to the duration of the illness, and the
predominance of manic or depressive phases. Thus, the longer the duration of
the illness, the greater the possibility of mixed states. Additionally, the
HABILOS showed with regard to the ratio of manic to depressive episodes
that the more manic the course, the greater the possibility of mixed states (see
Chapter 9). The frequency of mixed states has been reported to be higher among
females; although this finding is controversial (Marneros et al., 1991a, b;
McElroy et al., 1992, 1995; Akiskal et al., 1998; Arnold et al., 2000). Some
studies suggested that mixed states are not uncommon in childhood and
adolescence (Geller and Luby, 1997; McElroy et al., 1997) (see Chapter 10).

The classical work of Kraepelin (1899, 1913, 1921) and of Weygandt (1899)
provided a rich and fascinating description of mixed states. Modern studies
confirm the observations of the classical literature that depressive symptoms are
common in mania and hypomania (Kotin and Goodwin, 1972), and vice versa:
manic features can also occur in depression (Himmelhoch, 1979; Koukopoulos
et al., 1992, 1995, 2000; Bauer et al., 1994; Perugi, et al., 1997; Akiskal et al., 1998;
Cassidy et al., 1998a, b; Dilsaver et al., 1999; see Chapter 7). Contemporary data-
based studies provide support for conceptualized mixed states broadly and dimen-
sionally, as well as categorically (McElroy et al., 2000). However, one has to be
aware that the broader the definition, the greater its shortcomings.
In addition to a mixture of manic and depressive symptoms, mixed states are
also frequently characterized by anxiety, suicidal tendencies, and catatonic and
psychotic symptoms (Kraepelin, 1899, 1913; Weygandt, 1899; McElroy et al., 2000;
21 Beyond major depression and euphoric mania

Kruger et al., 2003). In the classical descriptions of Kraepelin and Weygandt, as
well as the modern studies, such as Winokur et al. (1969), Post et al. (1989),
Cassidy et al. (1998a, b); Dilsaver et al. (1999), and Marneros et al. (see Chapter 9),
anxiety symptoms are not uncommon in mixed states. Anxiety symptoms appear
to correlate with depressive symptomatology (Kraepelin, 1899, 1913; Post et al.,
1989; Cassidy et al., 1998a, b; Dilsaver et al., 1999; see Chapter 9). Although
suicidal symptoms during mixed episodes clearly occur, their reported frequency
varies considerably “ between 55% (Dilsaver et al., 1994) and 14% (Marneros et al.,
1991a, b). Nevertheless, all available data show a considerably greater frequency
of suicidal symptoms in mixed states than in pure mania (Dilsaver et al., 1994:
55% versus 2%, Strakowski et al. (1996): 26% versus 7%, Marneros et al., 2004: 14%
versus 0% for pure manic disorder, and for schizoaffective mixed episodes
22% versus 1% pure schizomanic episodes). It should be noted that the investiga-
tions of Marneros et al. (1991a), in contrast to the other studies noted above, have
the advantage of being longitudinal, considering all episodes during a period of
more than 25 years. As noted by Kraepelin, psychotic symptoms are not uncom-
mon in mixed states. Nevertheless, the occurrence of psychotic symptoms, especially
mood-incongruent symptoms, gives rise to the question of differential diagnosis “
mixed bipolar episode or schizoaffective? Although DSM-IV and ICD-10 define
schizoaffective bipolar mixed episode, there is limited research on the topic
(Marneros, 1986“2004). Findings on schizoaffective mixed states are presented
in Chapter 8 of this book.
Catatonic symptoms can also occur in mixed states. Kruger et al. (2003) assert
that, in spite of the assumption that catatonic symptoms are associated with good
prognosis in psychotic or affective disorders, the opposite is true in the case of
mixed states: catatonic symptoms in mixed bipolar states are associated with
greater severity and poor prognosis.

Onset, course, and outcome
Studies on the onset, course, and outcome of mixed states are somewhat incon-
sistent. One of the problems in the literature are the terms ˜˜patients with mixed
mania™™ or ˜˜patients with mixed states.™™ Perhaps the correct formulation is
˜˜patients who have at least one mixed episode during their course.™™ That is, as
noted above, patients having mixed states usually also have pure depressive
and pure manic episodes during their course, and in some cases, also schizo-
depressive and schizomanic episodes. Perhaps it would be helpful if mixed states
were defined according to the predominance of mixed symptomatology over pure
manic or schizoaffective symptomatology (Marneros et al., 1991a, b; see Chapter 9).
The relationship between mixed states and age at onset is of interest. Some
studies have reported that patients with mixed states have a younger age at
22 A. Marneros and F. K. Goodwin

onset (Nunn, 1979; Post et al., 1989). Marneros et al. (Chapter 9) also found
this relationship, but noted that it only applied to schizoaffective mixed episodes.
On the other hand, some have found no differences in age at onset in patients
with or without mixed states (Marneros et al., 1991a, b; Perugi et al., 1997;
see Chapter 9), while one study actually found that patients with mixed
states had an older age at onset (Strakowski et al., 1996). McElroy et al.
(1997) found that adolescent manics were more likely to be mixed than adult
The duration of a mixed episode, as initially described by Kraepelin and
by Weygandt, is longer and more complicated than pure manic or depressive
episodes “ an observation replicated by some modern investigations (Keller et al.,
1986; Dell™Osso et al., 1991; Marneros et al., 1991a), but not all: Calabrese
and Delucchi (1990), for example, found mixed episodes to be shorter, while
Winokur et al. (1969) found them to be equal in length. In the Cologne study
(Marneros et al., 1991a), mixed episodes were longer than other episodes, but,
12 years later, the same team found no differences between mixed manic and
pure manic episodes. The authors note that the difference might be explained
by the fact that the population of the initial Cologne study had been treated only
very rarely with anticonvulsants like valproate or carbamazepine, but the popula-
tion of the later study quite frequently received anticonvulsant therapies (valpro-
ate, lamotrigine, carbamazepine). However this later study found that
schizoaffective manic episodes were significantly longer than any other kind of
episode (see Chapter 9).
The initial observations of Kraepelin (1899), and Weygandt (1899), that the
outcome of patients with mixed states is much more unfavorable, was replicated in
the Cologne study (Marneros et al., 1986“1991) and in the later HABILOS study,
as well as being noted in many contemporary studies (Himmelhoch et al., 1976b;
Keller et al., 1986; Prien et al., 1988; Cohen et al., 1988; Tohen et al., 1990; McElroy
et al., 1995; Perugi, et al., 1997). However, not all studies agree. Thus, Winokur
et al. (1969) and Keck et al. (1998) reported no difference in outcome between
patients with mixed versus patients with pure mania.

The comorbidity of mixed states with other psychiatric conditions is receiving
increasing attention (McElroy et al., 2000), but the findings are controversial
(Brieger and Marneros, 1999; see Chapter 12). Thus, one report noted higher
rates of comorbid substance abuse in patients with mixed states (Himmelhoch
et al., 1976a, b), but others did not find such an association (McElroy et al., 1995).
However, the latter study did note a higher rate of comorbid obsessive-compulsive
disorders (McElroy et al., 1995) (see Chapter 12).
23 Beyond major depression and euphoric mania

Family history
Few systematic data on the family history of patients with mixed states exist.
Perugi et al. (1997) did not report any differences in family history between
patients with mixed and patients with pure manic states.

Although the data regarding treatment of mixed states are also controversial, there
is a reasonable amount of data suggesting that lithium may be less effective in the
short- and possibly long-term treatment of mixed states than pure mania
(Goodwin and Jamison, 1990; McElroy et al., 2000). Valproate, lamotrigine, and
possibly atypical antipsychotics, especially clozapine and olanzapine, may be more
effective than lithium for patients with mixed episodes. However, the data are
based on studies using different definitions of mixed states, so we need further
comparative studies. Also, some studies suggest that antidepressant agents may
exacerbate mixed states (Koukopoulos et al., 2000, McElroy et al., 2000; see
Chapter 3).

Future perspectives on mixed states
As Perugi and Akiskal have pointed out (see Chapter 2), mixed state does not
represent a mere superimposition of affective symptoms of opposite polarity, but
a complex process of temperamental, affective, and other components “ mixed
states might be considered the most eloquent expression of a neurophysiological
Today, more than 100 years after the publication of the first book on mixed
states by Wilhelm Weygandt in 1899, our understanding of the condition has
increased, but there are still uncertainties and gaps. What we need is much more
research on the topic. What are the major issues in designing such research?
* First of all, we need a single consensus, which takes into account the advantages

and disadvantages of broader definitions versus more narrow definitions.
* How do we operationalize and assess the boundaries drawn by Kraepelin

between ˜˜transitional forms™™ (which should represent a stage in between
when depression changes to mania and vice versa) and ˜˜autonomous forms™™
(which should mean mixed disorder on its own)?
* Once a consensus definition is established, we should be able to clarify some of

the following points:
* What is the gender distribution?

* How stable are mixed states over the course of illness?

* Are the mixed states a challenge of the bipolar I versus bipolar II dichotomy

(a question raised by Vieta et al. in Chapter 4 of this book)?
24 A. Marneros and F. K. Goodwin

Fig. 1.9 Periodic mania with development in folie circulaire (periodische Manie mit Ausgang in
¨res Irresein: Kraepelin, 1913).

Rapid cycling
The term ˜˜rapid cycling™™ is a modern one. However, the phenomenon of frequent,
or very frequent, recurrence of manic-depressive and mixed episodes was very well
known early in the evolution of scientific psychiatry. Emil Kraepelin was perhaps the
first who systematically described the phenomenon of rapid cycling (1899, 1913).
Of course, Kraepelin never used the term ˜˜rapid cycling™™ (Figs. 1.9“1.11). In one
of the earliest uses of the method of retrospective and prospective chart review,
Kraepelin documented the frequency and duration of episodes in life charts; he
described patients with more than four episodes per year, those with many more
than four episodes, patients with very short symptom-free intervals, and those
with no free intervals at all (Kraepelin, 1913). However, in the following decades,
essentially nothing more was done.
The term ˜˜rapid cycling,™™ as well as the increasing interest in this phenomenon,
also grew out of the psychopharmacological revolution. Dunner and Fieve first
coined the term ˜˜rapid cycling™™ in 1974, in what Calabrese et al. called a ˜˜landmark
paper™™ (Calabrese et al., 2000), which summarized longitudinal data designed to
evaluate clinical factors associated with lithium prophylaxis failure. But the
25 Beyond major depression and euphoric mania

¨ßiges, fast das ganze Leben ausfu
Fig. 1.10 Irregular almost lifelong folie circulaire (unregelma
¨res Irresein: Kraepelin, 1913).

¨res Irresein mit einleitenden
Fig. 1.11 Folie circulaire with beginning psychotic attacks (zirkula
¨llen: Kraepelin, 1913).
deliranten Anfa
26 A. Marneros and F. K. Goodwin

boundaries between ˜˜rapid cycling™™ (having at least four episodes in a year) and
˜˜not rapid cycling™™ (having fewer than four episodes per year) are in fact arbitrary,
although Dunner and Fieve found that most lithium non-responders belonged to
the group of more than four episodes.
The subsequent work of Wehr and Goodwin (1979) replicated and extended the
rapid-cycling findings of Dunner and Fieve, and additionally proposed that anti-
depressant agents could contribute to the manifestation of rapid cycling. This
finding was also later replicated (Calabrese et al., 1991, 1993; see Chapter 3).
Calabrese et al. (2000) pointed out:

The DSM-IV definition of rapid cycling describes it as a course modifier and is predicated for the
most part on the Dunner and Fieve conceptualization of the phenomenon:


. 5
( 68 .)