. 56
( 68 .)


mania, presumably by stimulating postsynaptic receptors. Paradoxically, single
doses of amfetamine may improve mania in some cases (Chiarello and Cole, 1987).
The onset of the first symptoms of bipolar disorder is frequently in adolescence,
with the emergence of subsyndromal mood swings or cyclothymia. The diagnosis
of bipolar disorder is commonly delayed by as much as 10 years. Recognition of
these prodromal features is particularly important in young people with a family
331 The treatment of bipolar mixed states

history of major affective disorder. Recently there have been more frequent reports
of mania in children (Geller et al., 2001), a phenomenon that has been suggested to
be linked to the use of stimulant and antidepressant drugs in young people
thought to have attention-deficit hyperactivity disorder or depression, respectively
(Reichart et al., 2000). The age at presentation with bipolar disorder is earlier in
those who had previous exposure to methylphenidate than those who had not
(Shulman et al., 2002).
All psychiatrists should now be familiar with the problems of cannabis in their
patients, and in future may be able to intervene more effectively if a cannabinoid
antagonist becomes available. However the most serious immediate challenge is
from ˜˜crack™™ cocaine, which is short-acting and highly addictive, and dramatically
destabilizes both the illness and social structure. In urban areas with heavy trade in
this substance, it is linked to violent gangs, and patients put their health and safety
at risk. A ˜˜dual diagnosis service™™ led by a psychiatrist with a special interest in
drug misuse may be needed to address the complex psychiatric, medical, and legal
problems that arise; a specialized ward area must have sufficient security to exclude
drug dealers and to treat the patients in a state free of harmful substances.
Keck et al. (1998) observed 134 consecutive manic patients for 12 months after
discharge, 58 having had mixed episodes according to DSM-IIIR (American
Psychiatric Association, 1987). They found no significant difference between the
two groups in any outcome variable, including alcohol or drug misuse. Substance
misuse was however associated with poor compliance with treatment.

Modified by organic brain disease
The presentation of mania may also be modified, leading to mixed forms, by the
comorbid occurrence of organic brain disease (Himmelhoch and Garfinkel, 1986).
A total of 31/37 patients with mixed mania had some indication of possible
neuropsychiatric impairment (seizures, paroxysmal electroencephalogranus,
migraine, a history of severe head injury or perinatal insults, or hyperactivity).
These were postulated to be a correlate of both mixed mania and of non-response
to lithium. By contrast, examination of children who became manic subsequently
has revealed a superior level of motor development and intelligence than controls
and no greater risk of obstetric complications (Cannon et al., 2002). A recent
addition to this controversy is the discovery on magnetic resonance imaging of
deep white-matter lesions, particularly in the frontal and parietal brain areas of
patients with major depression, including bipolar disorder, in excess of those seen
in psychiatric controls. These lesions may be associated with vascular ischemic
pathology (Thomas et al., 2002), and their presence is associated with treatment
resistance (Moore et al., 2001). A possible association with cognitive impairment is
332 J. Cookson and S. Ghalib

being investigated, but as yet no clear phenomenological association with mixed
states has been reported.
It has been reported that mania occurring in later life (after the age of 65) is
more likely to present in mixed episodes (Shulman and Post, 1980), but this has
not been confirmed. The situation in people with a first presentation of mania in
older age is being researched in detail. A study based on Danish national case
registers found that the diagnosis of dementia was associated with an increased risk
of both depression and mania over subsequent years compared with control
groups with osteoarthritis or diabetes (Nilsson et al., 2002).

Ultrarapid cycling
Operationally mixed states may be defined in terms of the coexistence of symp-
toms of both mania and depression in sufficient numbers or at sufficient levels of
severity. It should be noted that DSM-IV and Tenth Revision of International
Classification of Disease (ICD-10: World Health Organization, 1993) differ some-
what in their definitions of mixed states. ICD-10 includes patients showing a
mixture or alternation of manic and depressive symptoms for 2 weeks. This
would include people who might otherwise be classified as having rapid-cycling
bipolar disorder. DSM-IV requires full mania and full depression for 1 week.
Rapid cycling is included as a course specifier of bipolar disorders in DSM-IV
(American Psychiatric Association, 1994), and defined as ˜˜at least four episodes of
a mood disturbance in the previous 12 months that meet criteria for manic
episode, a hypomanic episode, or a major depressive episode.™™ Episodes are
demarcated by either partial or full remission for at least 2 months, or a switch
to a mood state of opposite polarity.
Ultrarapid cycling describes four or more episodes a month. Recently, ultradian
cycling has been described in otherwise typical bipolar patients; mood changes in
such patients occur in a matter of minutes or hours (Kramlinger and Post, 1996).
Perugi et al. (2000), reporting on 320 bipolar patients, have shown that bipolar
illness with depression at onset is significantly more likely than manic and mixed
states onset to develop rapid cycling, suicidal behavior, and psychotic symptoms.
This study confirmed that rapid cycling was distinct from mixed states. However,
in ultrarapid cycling ( 4 episodes/month) there may be considerable overlap with
mixed states.
Patients with ultradian cycling can also be found among those with a diagnosis
of borderline personality disorder, but these also present with extensive difficulties
in maintaining relationships. Their treatment poses unusual difficulties.
Antidepressant response to valproate in rapid-cycling patients was associated
with absence of borderline personality disorder (Calabrese et al., 1993).
333 The treatment of bipolar mixed states

The treatment of rapid cycling is reviewed elsewhere (see Chapter 3) and will not
be considered further here.

With mood-incongruent psychotic features
Patients with affective disorder in whom the content of psychotic symptoms is
incongruent with the prevailing mood are sometimes diagnosed as having mixed
(bipolar) states (Kraepelin, 1921). Mixed episodes of bipolar disorder are notor-
ious for signs and symptoms of protracted psychotic disorganization. Confusion
and psychotic features, including mood incongruence, are also common clinical
presentations (Perugi et al., 1997; Dell™Osso et al., 1991, 1993).
Dell™Osso et al. (1991) found higher rates of mood incongruence in the psy-
chotic features of mixed as opposed to pure manic patients using DSM-IIIR
criteria. However the inclusion of such patients in the bipolar grouping rather
than schizoaffective or schizophrenic broadens the diagnosis to an extent that
may not be helpful when attempting to relate diagnosis to response to specific
treatments. For example, Prien et al. (1972) found that antimanic response to
lithium was less in mania accompanied by severe symptoms or psychotic features;
this is discussed by Goodwin and Jamison (1990) and Goodnick and Meltzer
(1984). Concurrent psychotic features are associated with a high rate of recurrence
with lithium-based prophylaxis (Tohen et al., 1992).
On the other hand catatonic symptoms are relatively common in severe mania;
they were present in three of 60 patients with pure mania and 24 of 39 with mixed
mania (Braunig et al., 2002). Although regarded as ˜˜mood-incongruent™™ in DSM-III,
they became a specifier in DSM-IV.
This form will not be discussed further here.

Schizoaffective mixed states
It might be more appropriate to define mood-incongruent psychotic features
more clearly and to separate those with incongruent content into the schizoaffec-
tive class that has been described by Marneros and Angst (2000). The treatment of
this form will not be discussed here. It should be noted that validity is given to the
inclusion of such patients in the bipolar group by genetic and twin studies,
showing the diagnosis in identical twins or triplets, where one member has typical
mania (McGuffin et al., 1982).

Clinical trials in mania and mixed states
Evidence about drug treatment of mixed states arising from clinical trials should
recall the nature and limitations of such trials. These are summarized in Table 15.2.
334 J. Cookson and S. Ghalib

Table 15.2 Limitations of clinical trials of drugs in mixed states

1. The studies are powered to achieve significance when all subtypes of mania are included
2. Most recent large-scale studies showed a particularly high placebo response
3. Active comparators are often used at a single dose that may not be optimal
4. Patients giving informed consent to placebo-controlled trials may not be representative
5. Inclusion of treatment-resistant patients may bias results
6. Acute studies have a short duration
7. Abrupt withdrawal from previous medication may affect the clinical condition

First, the recent trials have mainly been carried out in mania rather than depres-
sion. They are not intended to study subtypes of mania, and are not ˜˜powered™™ to
do so. Subanalyses of groups with or without psychotic features or with depressive
symptoms are generally underpowered to answer the questions in those subgroups.
That is, they do not necessarily include sufficient numbers of patients in mixed states.
Most recent studies showed a high placebo response, as judged by 50%
improvement in the baseline severity of the mania rating scale, for example, 25%
(Bowden et al., 1994), 24% (Tohen et al., 1999) and 43% (Tohen et al., 2000) in
monotherapy studies. In adjunctive studies similar high rates have been found:
45% (Tohen et al., 2002a), 42% (Yatham 2000) and 46% (Muller-Oerlinghausen
et al., 2000). The only exception was a placebo response rate of 9% in the trial of
valproate in lithium-resistant mania with recruitment of inpatients (Pope et al.,
1991). These high placebo response rates may occur for a variety reasons, but they
are not typical of what is expected in acute mania, a condition that is notorious for
the difficulties it can pose in nursing management unless effective medication is
given. This suggests that the findings may not be generalizable to typical patients
encountered in routine clinical practice. The high placebo response rate also
reduces the effect size that the active drug is able to achieve, and makes it more
difficult to identify predictors of response.
Active comparators such as haloperidol tend to be used at a single dose that may
not be optimal.
Patients able to give informed consent to placebo-controlled trials may not be
representative of other patients with mania (Licht, 2002). In general, they are likely
to be less severely ill and to have fewer psychotic features and less aggressive behavior
(Licht et al., 1997). People under compulsory detention and those using illicit
drugs are also mostly excluded.
Patients available for trials may be more likely to be resistant to previously
available treatments. In a trial of valproate in acute mania, 42% had a history of
non-response to lithium (Bowden et al., 1994). This would clearly bias the results
in favor of valproate rather than lithium.
335 The treatment of bipolar mixed states

Table 15.3 Prediction of response to antimanic drug in models

Model Prediction

1. Mixture of elements (mood,
Partial response (in manic elements)
activity, thinking)
2. Severe stage of mania Proportionately large response
3. Dysphoric mania Different from response in pure mania
4. Depression as characterological
Depression improves as mania improves
response to mania
5. Manic defense in depression Depression worsened
6. Transition state during a cycle MDI: shortened mania; less severe depression;
(MDI/DMI) stabilize mood ˜˜from above ™™
7. Predominantly depressed
Depression persists or worsens
(BP-II, Dm)
8. Mania modified by substance
Resistance while comorbidity persists
9. Modified by organic brain disease Resistance while comorbidity persists

See text for definition of abbrevations.

The relatively short duration of studies in acute mania (3“4 weeks) means that
longer-term outcomes, and in particular the switch into depression, may not be
explored in the duration of the trial.
Discontinuation of lithium over a period of less than 2 weeks can increase the
likelihood of mania developing or worsening. Abrupt discontinuation can lead to
the development of mania in as many as 50% of bipolar patients who have been
stabilized on the drug (Mander and Loudon, 1988; see Cookson, 1997). Little is
known of how the discontinuation of other drugs may affect the clinical picture
and response to treatment.

Treatment responses in different models of mixed states
The different concepts or ˜˜models™™ of the nature of mixed states lead to different
predictions as to how drugs are likely to affect the condition. The predictions with
respect to treatment with a drug that has antimanic but not antidepressant
properties are shown in Table 15.3.
In general, with the exception of model 2 (severe stage), the models predict that
mixed states will respond less fully to antimanic drugs, such as antipsychotics.
There is limited evidence that dysphoric mania (model 3) responds differently,
and that in severe mania there is a proportionately greater response (model 1).
Thus, in a double-blind placebo-controlled cross-over study of carbamazepine in
336 J. Cookson and S. Ghalib

19 acutely manic patients, the 12 who responded to carbamazepine had higher
scores of severity of mania and tended to be more dysphoric (Post et al., 1987).

Antipsychotics in mania
The response to antipsychotic medication in mania has recently been subject to
careful investigation in clinical trials of atypical antipsychotics, particularly olanzapine
and risperidone. These studies are reviewed in detail in Chapter 16. Two studies
included, as a comparator, the classical antipsychotic haloperidol which is the
most widely used antipsychotic for mania (see Cookson, 2001). All three drugs
(haloperidol, olanzapine, and risperidone) tended to produce as great an improve-
ment in mania ratings in the more severe and psychotic groups of patients as in less
severe non-psychotic patients (in keeping with model 2).

Cortisol levels during response to antipsychotics
Elevated serum cortisol levels are found in mania. This elevation is correlated
significantly with the severity of mania (Cookson et al., 1985a), but is particularly
high in mixed states (Swann et al., 1992). Evans and Nemeroff (1983) found that
mixed manics showed more resistance of plasma cortisol to suppression by
dexamethasone than pure manics, although Swann et al. (1992) found cortisol
non-suppression in both pure and mixed mania. Krishnan et al. (1983) found
non-suppression in all of 10 consecutive patients with mixed states.
During treatment with pimozide, cortisol levels gradually return towards nor-
mal, with a time course similar to that of clinical improvement (Cookson, 1985).
By contrast, during treatment with haloperidol, there appears to be a dissociation
between an early normalization of cortisol levels within 3 days and a more gradual
clinical improvement during 2 weeks of treatment (Cookson et al., 1985b). This
apparent difference between haloperidol and pimozide might be related to the
different pharmacology of the two drugs. Both drugs block dopamine receptors,
but haloperidol in addition blocks norepinephrine alpha1-receptors in humans
(Szabadi et al., 1981). Alpha1-receptors are known to be involved in the control of
cortisol secretion (Rees et al., 1970). Alpha1-receptor blockade is thought to
contribute to the sedative effects of antipsychotics (Peroutka and Snyder, 1980).
It may account for the early transient sedative effects seen in mania with haloperidol
(Cookson et al., 1983). Levels of norepinephrine in the CSF are raised in mania (Post


. 56
( 68 .)