<<

. 60
( 68 .)



>>

psychotic agents for bipolar disorder. The authors of this chapter have contributed
to the large portfolio of studies of olanzapine in bipolar disorder, several of which
yield information on treatment response of rapid-cycling patients and those with
psychotic or dysphoric mania. At this writing, data comparable in breadth or
depth to those on olanzapine are not yet available for other atypical antipsychotic
agents. However, research is ongoing and one or more controlled acute mania
studies have been encouraged for risperidone (Hirschfeld et al., 2004; Sachs et al.,
2002a, 2002b; Vieta et al., 2002), ziprasidone (Keck et al., 2003a), quetiapine
(Sachs et al., 2004), and aripiprazole (Keck et al., 2003b). Data are not available
for other potentially useful atypical agents currently under development, such as
iloperidone.

Method
We briefly review available data on antipsychotic agents for the treatment of
patients with bipolar disorder. The review is limited to controlled, double-blind
clinical trials; these were identified through Medline search (as of March 2003);
cross“referencing bibliographies of identified manuscripts; olanzapine trials con-
ducted by us; and preliminary reports obtained at public scientific meetings
through 2002. This fourth method offers up-to-date information in this fast-
developing field, but carries significant limitations, such as less detailed presenta-
tion of methods and/or findings than is typical in manuscripts; incomplete peer
review; and the likelihood that some relevant presentations or posters were not
available to us.
356 R. W. Baker et al.


Our primary focus is the relative response within diagnostic subgroups, espe-
cially psychotic versus non-psychotic, mixed versus manic, and rapid versus non-
rapid cycling. When available, we report the group-by-treatment interaction,
because a significant interaction would signal the potential usefulness of diagnostic
categorization to treatment selection. Most data are from mania studies. In order
to maintain consistent presentation across studies, we focus on improvement in
the primary mania rating scale; response rates; and relative results for diagnostic
subgroups. Further, as older antipsychotic agents are constrained in use for bipolar
disorder by neurological and depressogenic side-effects, also reviewed are reports
of: study completion rates; impact of treatment on depression ratings; and extra-
pyramidal adverse events. In the interest of comparison across studies, we generally
convert raw change to percentage change. For example, percent improvement
in mania is calculated by dividing mean within-treatment group improvement by
the mean baseline mania rating score for that treatment group. P-values are not
revised with the conversion to percentage change, but are those reported for mean
change and interaction comparisons in the primary studies.

Clozapine
Clozapine was the first of the atypical antipsychotic agents, with clinical trials in
schizophrenia starting over three decades ago. Nevertheless, there are no blinded
studies of its use in bipolar disorder. Small, open trials are encouraging, including
use adjunctive to other mood stabilizers for up to 1 year in a mixed group of
schizoaffective and bipolar patients (Suppes et al., 1999). A number of open
reports (total n ¼ 42) on its efficacy for treatment“refractory mania (Calabrese
et al., 1996; Green et al., 2000) have been reported. Controlled double-blind
studies in this subgroup would be welcome.

Risperidone
Two available double-blind trials of risperidone in patients with bipolar disorder
explored its usefulness as an adjunct to mood stabilizers for treating acute mania.
One of these studies was positive. Sachs and collaborators (2002ab) reported a
double-blind 3-week study in patients with acute mania or mixed episodes being
treated with divalproex or lithium, who were randomly assigned to adjunctive
treatment with risperidone 1“6 mg/day (n ¼ 52), haloperidol 2“12 mg/day
(n ¼ 53), or placebo (n ¼ 51). Improvement from baseline to endpoint in
Young Mania Rating Scale (YMRS) was superior among patients receiving risper-
idone (51%) or haloperidol (49%) compared to placebo (29%). Though group-
by-treatment interaction analysis is not reported, the mania rating mean change
results suggest consistent therapeutic effects across psychotic and non-psychotic
subgroups for both haloperidol and risperidone as adjunctive treatment to mood
357 Use of atypical antipsychotic agents


stabilizers. Interestingly, however, results were not consistent across mixed and
manic episodes; both adjunctive antipsychotic treatments were effective for manic
patients and neither was for mixed patients. In fact, among mixed patients, the
group who had placebo added to lithium or divalproex had the greatest mean
improvement. Depression ratings were not reported. Discontinuation rates during
the 3-week study were 35% on risperidone, 53% on haloperidol, and 49% on
placebo. Extrapyramidal symptom ratings worsened significantly in the haloper-
idol group, but not the risperidone group, compared to placebo. Adverse events of
extrapyramidal disorder were reported in 28% of patients receiving adjunctive
haloperidol compared to 13% of those receiving risperidone and 4% of those on
placebo.
A second study enrolling 150 patients in acute manic or mixed episodes failed to
confirm the superiority of risperidone over placebo as an adjunct to divalproex,
lithium, or carbamazepine (Yatham, 2000). In this study, the contribution of
risperidone was least apparent among those patients using carbamazepine as
their primary mood stabilizer, raising the possibility that accelerated metabolism
secondary to activation of cytochrome CYP-2D6 may have been an important
contributor to the failure of risperidone to differentiate from placebo. Detailed
information on efficacy within diagnostic subgroups is not yet available for
this study.
A South African group (Segal et al., 1998) published a double-blind, random-
ized monotherapy comparison of risperidone 6 mg/day and haloperidol 10 mg/day to
lithium (mean level 0.53 mmol/l week 1, 0.62 mmol/l week 3, 0.72 mmol/l week 4)
for hospitalized patients with acute mania, many with psychosis. Improvement
from baseline YMRS was not statistically significantly different among the treat-
ment groups, with 55% reduction on lithium, 43% on risperidone, and 41% on
haloperidol. Depression ratings are not reported. However, extrapyramidal symp-
toms were similar in the risperidone and haloperidol groups, with both worsening
substantially compared to the lithium group. Given dose-related extrapyramidal
side-effects with risperidone, it is possible that lower doses would have been
associated with lower neurological side-effects than haloperidol 10 mg; likewise
the small sample size (15 patients per group) and absence of a placebo control
necessitate caution in interpreting the lack of statistically significant difference in
mania improvement across these groups.
Two placebo-controlled trials have shown promising efficacy of risperidone
monotherapy for acute mania. In a USA-based study, Hirschfeld and collaborators
(2004) evaluated risperidone 1“6 mg/day (mean modal dose 4.1 mg/day) for
patients hospitalized with acute mania. Mean YMRS improvement after 3 weeks
of treatment was superior on risperidone (38% improvement, n ¼ 134) than on
placebo (17% improvement, n ¼ 125, P<0.001). Response rates, defined as at least
358 R. W. Baker et al.


50% reduction in YMRS, were 43% on risperidone and 24% on placebo (P<0.01).
Efficacy was significant for both psychotic and non-psychotic patients; though
group-by-treatment interaction analysis is not provided, the efficacy of risperi-
done appears roughly comparable in both groups (advantage over placebo in
mean YMRS change of 6.3 points in psychotic patients versus 5.3 points in non-
psychotic patients). Unfortunately this trial does not address other subgroups:
patients with a history of rapid cycling and/or current mixed episode were not
enrolled. Changes in depression ratings were minimal in both groups and did not
differ significantly between risperidone and placebo. Drop-out rates for the 3-week
trial were 44% on risperidone and 58% on placebo (P-value not provided). Sixteen
percent of risperidone-treated patients had adverse events of ˜˜hyperkinesia™™ versus
5% on placebo (P-value not provided).
Vieta and colleagues (2002) reported a 3-week inpatient study of risperidone
versus placebo for acute manic or mixed episodes. At doses of 1“6 mg/day,
risperidone (n ¼ 146, mean modal dose 5.6 mg/day) was associated with superior
YMRS improvement than placebo (n ¼ 144). Treatment response, defined as 50%
or more reduction in YMRS ratings, was achieved twice as frequently on risperi-
done (73%) as on placebo (36%, P<0.001). Changes in depression ratings are not
reported, though the authors indicate that their findings do not suggest induction
of depression with risperidone. Again, efficacy was significant in both psychotic
(approximately 60% of the overall sample) and non-psychotic subgroups. Relative
benefits to mixed versus classic manic episodes are not reported, and the impact
for rapid-cycling patients cannot be addressed as they were excluded from the
study. Three-week completion rates were relatively high “ 89% and 71% in risperi-
done and placebo groups respectively (between-group P-value not provided). Use of
anticholinergic medications and extrapyramidal symptom ratings are not described.
The rates of adverse events of extrapyramidal disorder in this 3-week study were
35% among risperidone-treated patients and 6% on placebo; tremor was an adverse
event for 10% of patients on risperidone versus 1% on placebo.

Olanzapine
Olanzapine is now among the best-studied agents for bipolar disorder, with over
1000 patients studied in double-blind mania studies, and recently reported, large
double-blind studies in both acute bipolar depression and bipolar relapse
prevention.
At least seven double-blind studies have explored the use of olanzapine for acute
mania, with superior efficacy observed in all three placebo-controlled studies
(Tohen et al., 1999, 2000, 2002b) and a valproate-controlled study (Tohen et al.,
2002a); statistically significant differences in acute mania improvement were not
found in three other active comparator-controlled studies (Berk et al., 1999;
359 Use of atypical antipsychotic agents


Zajecka et al., 2002; Tohen et al., 2003a). Olanzapine has proven relatively free of
the depressive symptom worsening and extrapyramidal side-effects that limit the
usefulness of conventional antipsychotic agents.
Two placebo-controlled studies of olanzapine monotherapy for acute mania
have been published (Tohen et al., 1999, 2000) and one for addition to treatment
with valproate or lithium (Tohen et al., 2002b).
A 3-week study compared olanzapine 5“20 mg/day (n ¼ 70) to placebo (n ¼ 66)
for inpatients in acute manic or mixed episodes (Tohen et al., 1999). Olanzapine-
treated patients had 36% mean reduction in YMRS, superior to 18% mean
improvement on placebo. Response rates, defined as 50% or greater YMRS
improvement, were 48.6% for olanzapine-treated patients and 24.2% for placebo-
treated patients. Efficacy was consistent across psychotic and non-psychotic
subgroups: interaction P ¼ 0.880. Interaction was also non-significant (P ¼ 0.998)
for manic versus mixed subgroups. In this case, the advantage of olanzapine over
placebo was evident for the manic group (37% versus 16% YMRS improvement),
but in the mixed group YMRS reduction on olanzapine averaged 28% versus 25%
on placebo. Interaction was likely non-significant because so few mixed patients
were enrolled in this study: there were only 12 per treatment group. Mean
Hamilton Depression ratings improved similarly in olanzapine (23%) and placebo
(21%) groups. The 39% discontinuation rate in the olanzapine group was sig-
nificantly lower than the 65% discontinuation rate on placebo. Change in extra-
pyramidal symptom ratings (Simpson-Angus and Barnes Akathisia scales) did not
differ between treatment groups.
A second trial (Tohen et al., 2000), this time of 4 weeks™ duration, compared
olanzapine 5“20 mg/day (n ¼ 55) to placebo (n ¼ 60) for inpatients in acute manic
or mixed episodes. The 51% mean reduction in YMRS during olanzapine treat-
ment was superior to the 28% mean improvement on placebo. Response rates,
defined as an individual™s mania rating improvement of 50% or more, were 65%
on olanzapine versus 43% on placebo (P ¼ 0.02). In this study, olanzapine™s
advantage was particularly evident in psychotic patients (mean YMRS reduction
of 52% versus 13% on placebo; in non-psychotic patients, mean improvements
were 51% and 44%). The test of interaction was nearing significance (P ¼ 0.110);
this finding is discrepant from other available studies, which point to olanzapine
efficacy as great or greater among non-psychotic than psychotic patients; it may
reflect the particularly large improvement on placebo for non-psychotic indivi-
duals in this particular trial. In this study, the efficacy of olanzapine was consistent
across mixed and manic episodes (interaction P ¼ 0.706). Mean Hamilton
Depression ratings trended to superior improvement on olanzapine (45%) com-
pared to placebo (28%). The 32% discontinuation rate in the olanzapine group
was significantly lower than the 58% discontinuation rate on placebo. Change in
360 R. W. Baker et al.


extrapyramidal symptom ratings (Simpson-Angus and Barnes Akathisia scales)
did not differ between treatment groups.
A 6-week trial (Tohen et al., 2002b), predominantly in outpatients with at least
moderate symptoms of a manic or mixed episode, compared double-blind addition
of olanzapine 5“20 mg/day (n ¼ 229) or placebo (n ¼ 115) to open treatment with
valproate or lithium that had been established for at least 2 weeks. Mean YMRS
improvement at endpoint was superior in the olanzapine group (59% reduction)
compared to the placebo group (40%). Response rates were 68% in the olanzapine
co-therapy group, versus 45% in patients receiving lithium or divalproex mono-
therapy (P<0.001). Antimanic efficacy was consistent across diagnostic subgroups:
psychotic versus non-psychotic (interaction P ¼ 0.440) and mixed versus manic
(interaction P ¼ 0.708). Improvement in Hamilton Depression ratings was also
superior on olanzapine (34% reduction) compared to placebo (7%). Interestingly,
this significant antidepressant effect largely reflected an improvement in mixed
patients with high baseline depression scores.
Controlled trials of mood stabilizers generally have not addressed the simulta-
neous improvement of manic and depressive features seen on olanzapine in this
study. Thirty percent of the olanzapine group and 29% of the placebo group
discontinued treatment during the 6-week study. Change in extrapyramidal
symptom ratings (Simpson-Angus and Barnes Akathisia scales) did not differ
between olanzapine and placebo groups, though treatment-emergent tremor was
more common on olanzapine (23%) than placebo (13%).
A South African group (Berk et al., 1999) conducted a 4-week double-blind
comparison of olanzapine to lithium in small cohorts of hospitalized patients with
acute mania (n ¼ 15 for each group). Though olanzapine trended to superiority in
global ratings, change in mania ratings and extrapyramidal symptoms did not
differ between the groups, and depression ratings are not reported. As with their
similar report (Segal et al., 1998) comparing risperidone and lithium, the small
sample sizes limit conclusions from these results.
Olanzapine 5“20 mg/day (n ¼ 234) and haloperidol 3“15 mg/day (n ¼ 219) were
compared in a 12-week monotherapy study in patients with acute mania (Tohen
et al., 2003a). The primary outcome was remission from both mania and depression
at week 6, which did not differ significantly between groups (57% of olanzapine-
treated patients achieved remission, versus 46% on haloperidol). Importantly,
however, an interaction was found (P ¼ 0.09) between outcome and psychosis
status, suggesting that haloperidol and olanzapine are similarly effective for the
acute treatment of psychotic mania (remission rate of 49% in each group) whereas
olanzapine was more effective for non-psychotic patients (remission rate of 57%
versus 42% on haloperidol, P ¼ 0.04). There was no interaction (P ¼ 0.33) based on
diagnosis of mixed versus manic episode. Mean YMRS improvement at 6 weeks did
361 Use of atypical antipsychotic agents


not differ significantly between groups (69% on olanzapine and 77% on haloper-
idol), whereas Hamilton Depression rating improvement trended to superiority in
the olanzapine group (35% reduction versus 23% in the haloperidol group).
Discontinuation rates in the olanzapine group were 29% at week 6 and 40% by
week 12, compared to 36% and 47% respectively, in the haloperidol group. This
study is an interesting reminder that typical antipsychotic medications can be quite
effective for acute mania, perhaps especially for psychotic mania. However, there
was confirmation of the principal limitations of conventional agents, with switch to
depression occurring significantly sooner and in more patients in the haloperidol
group, and significantly greater extrapyramidal symptoms on haloperidol, as mea-
sured by Simpson-Angus and Barnes Akathisia, and Abnormal Involuntary
Movement scales. Several treatment-emergent extrapyramidal adverse events were
significantly more common on haloperidol than olanzapine, including akathisia
(30% versus 6%), extrapyramidal syndrome (24% versus 2%), hypertonia (18%
versus 5%), tremor (16% versus 6%), dystonia (7% versus 1%), dyskinesia (3%
versus 0%), and tardive dyskinesia (3% versus 0%). Moreover, switch to depressive
episodes trended to being more frequent among haloperidol-treated patients (17%)
than olanzapine-treated patients (9%), P ¼ 0.098, and time to switch to depression
was earlier in the haloperidol group (P ¼ 0.04).
Finally, two studies compared olanzapine to valproate monotherapy for inpatients
with acute mania. A study conducted by Zajecka and collaborators compared olan-
zapine (starting dose 10 mg/day, maximum 20 mg/day, n ¼ 57) to valproate (starting
dose 20 mg/kg/per day, maximum increase 1000 mg/day, n ¼ 63) for inpatients with
acute mania (Zajecka et al., 2002). Mean YMRS improvement at 3 weeks did not
differ significantly between olanzapine (53%) and valproate (48%). Response rates
and diagnostic subgroup results were not reported. Hamilton Depression ratings also
improved similarly in the two groups: 46% on olanzapine and 42% on divalproex.
Three-week completion rates were 68% on olanzapine, 62% on valproate, P-value
was not provided. Extrapyramidal symptom ratings on the Simpson-Angus or Barnes

<<

. 60
( 68 .)



>>