<<

. 13
( 68 .)



>>

relapse rates of children with a prepubertal and early adolescent bipolar disorder
phenotype. Of 93 subjects seen at baseline, 89 were seen at year 1 (95.7% reten-
tion). The rate of recovery from mania was 37.1%, and the rate of relapse after
recovery was 38.3%. No covariates were significantly associated with recovery or
relapse. The low recovery and high relapse rates supported the study hypothesis of
poor outcomes, which was made on the basis of similarity between the character-
istics of the prepubertal and early adolescent bipolar disorder phenotype (long
episode duration and high prevalence of mixed mania, psychosis, and rapid
cycling) and those of severe bipolar disorder in adults.
Additionally, Schraufnagel et al. (2001) reported the results of a study on the
patterns of affective illness presentation in children and adolescents, in relation to
pubertal maturation and family history. One hundred children/adolescents with
affective illness (72 boys and 28 girls; age range 2“20 years; mean age 10 years), who
were consecutively referred to the Pediatric Behavioral Neurology Program,
Children™s Medical Center at Dallas, Texas, USA, were evaluated for the pattern
and course of affective illness symptoms, family history, and pubertal stage. Seven
patterns of affective illness were identified. In the 65 prepubertal children (Tanner
stage 1), disorders with hypomanic/manic symptomatology were most common
(47 / 65, 72%), followed by dysthymia with bipolar features (18 / 65, 28%),
cyclothymia (11 / 65, 17%), hypomania (8 / 65, 12%), juvenile rapid-cycling bipolar
disorder/ultradian cycling bipolar disorder (8 / 65, 12%), and mania (2 / 65, 3%).
In contrast, the 26 fully pubertal adolescents (Tanner stages 3“5) had a predominance
of patterns with only depressive symptomatology (16 / 26, 61%), followed by
depression (12 / 26, 46%), along with juvenile rapid-cycling bipolar disorder/
ultradian cycling bipolar disorder (6 / 26, 23%), and dysthymia (4 / 26, 15%).
Affective illness, alcoholism, and drug abuse were prominent in the family
histories, regardless of the child™s pattern of symptoms. Family histories of character
disorder and Briquet™s syndrome were also common, but thought disorder, suicide,
and homicide were infrequent. This study supports the clinical observation that
the presentation of affective illness changes with age: manic features predominate
in younger children, whereas depressive symptomatology is more evident with
68 O. Elhaj and J. R. Calabrese


pubertal maturation. The results of these naturalistic studies warrant further
controlled and longitudinal studies to provide more generalizable conclusions
about the phenomenology of rapid-cycling bipolar disorder in children, and its
management.
In Calabrese et al.™s paper (2000a) contained in the preceding book, we reported
the following with regard to comorbidity: extensive literature exists on the pre-
sence of thyroid dysfunction in patients with bipolar rapid cycling. Some studies
do (Cho et al., 1979; Cowdry et al., 1983; Bauer et al., 1990; Kusalic, 1992; McKeon
et al., 1992) but most do not (Joffe et al., 1988; Nurnberg et al., 1988; Wehr et al.,
1988; Bartalena et al., 1990; Coryell et al., 1992; Shen, 1992; Cole et al., 1993; Maj
et al., 1994; Oomen et al., 1996; Post et al., 1997) suggest that rapid cycling is
associated with an underlying thyroid abnormality. Usually (Khouzam et al.
1991), the observed abnormality of thyroid dysfunction has been in the direction
of decreased end-organ function. Herz (1964) first proposed that rhythmic disorders
of mood might be caused by removal of the thyroid gland. Twenty-two recently
thyroidectomized patients were examined for evidence of psychiatric complications
in Frederiksberg Hospital, Copenhagen, Denmark. Ten exhibited postsurgical
psychiatric symptoms in the absence of any family psychiatric history. The authors
described this as the ˜˜endocrine psycho-syndrome™™ and specifically noted that eight
patients exhibited temporary attacks of depression soon after the surgery.
Cho et al. (1979) first demonstrated that the prevalence of lithium-induced
hypothyroidism was much higher in rapid cyclers (31%) than in non-rapid cyclers.
This was replicated by Cowdry et al. (1983), who noted overt hypothyroidism in
50.7% of 24 rapid cyclers and in none of 19 non-rapid cyclers. Elevated TSH levels
were present in 92% of the rapid cyclers and 32% of the non-rapid cyclers. Five years
later, the same group (Wehr et al., 1988) refuted their earlier finding that thyroid
dysfunction was no more common in rapid cyclers than in non-rapid cyclers. Bauer
et al. (1990) have carried out the most thorough examination of thyroid function,
reporting a spectrum of thyroid abnormalities in rapid cycling. They have also begun
a systematic examination of the potential mood-stabilizing properties of thyroid
supplementation, when used in augmentation of conventional mood stabilizers. Of
30 patients with bipolar rapid cycling studied prospectively for the presence of
thyroid failure, 23% had grade I hypothyroidism (decreased FTIs with overt signs
and symptoms), 27% had grade II (normal FTI, elevated TSH, and a single sign/
symptom), and 10% had grade III (everything is normal, but there is an augmented
TSH response to thyroid-releasing hormone). A median and modal frequency of
24 episodes/year with a maximal frequency of two episodes per day suggests that
episode counting was done with criteria inconsistent with the DSM-IV.
It is clear that there is an increase in the prevalence of alcohol and drug abuse in
patients with bipolar disorder (Regier et al., 1990). Whether rapid cyclers have an
69 Rapid-cycling bipolar disorder


increased prevalence of alcohol and drug abuse comorbidity compared to non-
rapid cyclers has not been explored. Whether patients with bipolar disorder and
comorbid alcohol or drug abuse/dependence have an increased prevalence of rapid
cycling has likewise not been explored. However, preliminary data suggest that
bipolar patients with comorbid alcohol and/or drug abuse/dependence cycle
frequently, consistently experiencing twice as many lifetime hospitalizations
(Keller et al., 1986; Brady et al., 1991; Sonne et al., 1994; Haywood et al., 1995).
Other manifestations of comorbidity in rapid cyclers have not yet been system-
atically studied. However, anecdotal reports have associated the onset of rapid
cycling with neurologic events or states such as strokes (Berthier, 1992), sub-
arachnoid hemorrhages (Blackwell, 1991), and profound mental retardation with
periodic aggressive acting-out behavior (Glue, 1989; Lowry and Sovner, 1992).
After the publication of the above-mentioned book, Calabrese et al. (2001)
reported that comorbidity with alcohol, cannabis, and/or cocaine abuse or depen-
dence appeared to alter prognosis by increasing the prevalence of poor compli-
ance, not by directly affecting the spectrum of activity of combined treatment with
lithium and divalproex. In that study, Calabrese et al. evaluated the spectrum of
efficacy of combined treatment with lithium and divalproex in a cohort of 84
patients with rapid-cycling BP-I or BP-II disorder comorbid with a current history
of either abuse of or dependence on alcohol, cannabis, and/or cocaine. At the time
of study entry, 86% of patients were using alcohol, 45% cannabis, and 40%
cocaine. Of those using alcohol, 71% met DSM-IV criteria for dependence and
29% for abuse. Of those using cannabis, 24% met criteria for dependence, and
76% for abuse. Of those using cocaine, 65% met criteria for dependence and 35%
for abuse. The profile of lifetime abuse/dependence was alcohol/cannabis/cocaine
(42%), alcohol and cannabis (22%), alcohol alone (20%), alcohol and cocaine
(12%), cannabis and cocaine (1%), cannabis alone (1%), and cocaine alone (1%).
Physiologic dependence was present in 65% of those using alcohol, 8% of those
using cannabis, and 57% of those using cocaine. These data suggest that the
majority of alcohol and cocaine use, but not cannabis use, in rapid-cycling bipolar
disorder is accompanied by physiologic dependence. A more detailed discussion of
their outcome results will follow later in the chapter. Additionally, substance abuse
has been linked to increasing the already high risk of suicide in rapid-cycling
bipolar disorder (Nierenberg et al., 2001; Sachs et al., 2001).


Pharmacotherapy

Treatment recommendations
While the newly revised practice guidelines for the treatment of patients with
bipolar disorder, published by the American Psychiatric Association in 2002,
70 O. Elhaj and J. R. Calabrese


recommended that the initial treatment for patients who experience rapid cycling
should include lithium or valproate, lamotrigine has also been considered a first-
choice option. It was also advised that for many patients, combinations of medica-
tions are required (American Psychiatric Association, 2002).
Most researchers and clinicians have come to realize that the depressive phase of
bipolar disorder is probably more difficult to treat, especially in the rapid-cycling
¨
subtype (Ghaemi et al., 2000; Moller and Grunze, 2000; Perugi et al., 2000; Sachs
et al., 2000a; Calabrese et al., 2001, 2002). Nevertheless, the use of antidepressants
in treating the depressive phase of rapid-cycling bipolar disorder continues to
draw significant controversy. While most have warned about the likelihood of
antidepressants worsening the course of rapid-cycling bipolar disorder, and have
strongly cautioned against using them alone (Ghaemi et al., 2000, Perugi et al.,
2000; Sachs et al., 2000a, 2000b; Calabrese et al., 2001, American Psychiatric
Association, 2002), others have warned that the decreased use of antidepressants
¨
might increase the risk of suicide (Moller and Grunze, 2000), and that antidepres-
sant use along with a mood stabilizer constitutes a low risk for worsening rapid-
˜
cycling course (Amsterdam and Garcia-Espana, 2000; Sachs et al., 2000a).
˜
Amsterdam and Garcia-Espana (2000) reported results indicating that no epi-
sodes of drug-induced hypomania or rapid cycling were observed during 6 weeks of
venlafaxine monotherapy for depression in women with BP-II and unipolar major
depression. This study™s limitations included that it was retrospective in nature and
limited in patient number, that only BP-II women were included in this study, and it
is possible that efficacy and the manic switch rate might have differed if BP-I women
were included. However, it is noteworthy that the limitations of this study make it
very difficult to formulate any generalizable conclusions about the use of anti-
depressants in treating rapid-cycling bipolar disorder.
A respective chart review of outpatients with affective disorders (n ¼ 85, with
both bipolar and unipolar disorders) was conducted over a 1-year period (Ghaemi
et al., 2000). The results indicated that bipolar disorder was found to be misdiag-
nosed as unipolar depression in 37% of patients who first saw a mental health
professional after their first manic/hypomanic episode. Antidepressants were used
earlier and more frequently than mood stabilizers, and 23% of this unselected
sample experienced a new or worsening rapid-cycling course attributable to
antidepressant use. These results suggest that bipolar disorder tends be misdiag-
nosed as unipolar major depressive disorder and that antidepressants seem to be
associated with a worsened course of bipolar illness. While the results of this study
are informative regarding the misdiagnosis of bipolar disorder and the eventual
negative consequences on its course, longer, more controlled studies with larger
sample size are still needed to clarify this long-lived controversy in the manage-
ment of bipolar disorder, and especially the rapid-cycling course.
71 Rapid-cycling bipolar disorder



Lamotrigine
Lamotrigine utilization as a first option in the treatment of rapid-cycling bipolar
disorder has gained more consensus (Sachs et al., 2000b; American Psychiatric
Association, 2002).
Calabrese et al. (2000b) reported on the first multicenter, double-blind, flexible-
dose, placebo-controlled, parallel-group study to examine the safety and efficacy
of lamotrigine for the long-term prophylaxis of mood episodes in patients with
rapid-cycling bipolar disorder.
A total of 324 patients meeting DSM-IV criteria for rapid-cycling disorder entered
the open-label phase, and 182 patients were eventually randomly assigned to the
double-blind maintenance phase. For patients entering the open stabilization phase,
the mean age was 38 years, the percentage of women was 59%, the percentage of
BP-I subtype was 69%, and the percentage of patients receiving thyroid supplements
for diagnosed hypothyroidism was 7%. At study entry, 57% of patients were
depressed, 20% were hypomanic or manic, 18% were euthymic, and 5% had mixed
states. The mean number of mood episodes in the 12 months prior to study entry was
6.3. The lifetime prevalence of psychosis was 27%, and the percentage of patients with
prior suicide attempts was 36%. Prior lifetime exposure to psychotropics included
lithium (68%), carbamazepine (27%), divalproex (57%), lamotrigine (< 1%), anti-
depressants (82%), and antipsychotics (27%). Concomitant psychiatric medications
at study entry included lithium (19%), carbamazepine (4%), divalproex (19%),
antidepressants (30%), and antipsychotics (7%). Although the most commonly
prescribed lifetime medications at the time of study entry were antidepressants, only
36% of patients reported having responded to them positively.
Lamotrigine was added to the patients™ current psychotropic regimens and
titrated to clinical effect during an open-label treatment phase. The treatment of
stabilized patients with other psychotropics was tapered off and randomly
assigned to lamotrigine or placebo monotherapy (in a 1:1 ratio) for 6 months
after being stratified for BP-I or BP-II disorder.
Lamotrigine dose was titrated in the 6-week preliminary phase to a target dose
of 200 mg/day. After week 5, lamotrigine dose increases were allowed in incre-
ments of 100 mg/week up to a maximum dose of 300 mg/day. In the randomized
phase, the double-blind medication dosage was also flexible and varied from 100 to
500 mg/day. The average daily dose was 288 Æ 94 mg.
Time to additional pharmacotherapy for emerging symptoms of a mood epi-
sode was the primary outcome measure. Secondary outcome measures included
survival in the study, percentage of patients stable without relapse for 6 months,
and changes in scores on the Global Assessment Scale (GAS) and the CGI-Severity
of Illness (CGI-S) scale.
72 O. Elhaj and J. R. Calabrese


Kaplan“Meier methodology was used to analyze survival data, and median
times to survival were calculated. Additionally, survival analyses were performed
for each bipolar subtype. The percentage of patients stable without relapse for
6 months was analyzed using the Cochran“Mantel“Haenszel chi-squared test.
Clinical efficacy scales (CGI-S, GAS) were evaluated using analysis of variance
(ANOVA) at an ¼ 0.05 level of significance using both observed and last-
observation-carried-forward (LOCF) data.
Treatment groups during the randomized phase (n ¼ 182) were similar with
respect to age, sex, race, medical history, psychiatric history, prior treatments,
response to treatments, and current psychiatric state. The majority of patients were
classified as having BP-I disorder (71%). A comparison of BP-I and BP-II patients
showed no differences on key parameters. Compared with BP-II patients, BP-I
patients had a greater prevalence of suicide attempts (40% versus 28%) and
average number of lifetime hospitalizations (2.3 versus 0.7).
Forty-nine placebo patients (56%) and 45 lamotrigine patients (50%) required
additional pharmacotherapy for emerging symptoms of a mood episode. The
difference between the two general treatment groups in time to additional pharma-
cotherapy did not achieve statistical significance. The median survival times
were 18 weeks for lamotrigine and 12 weeks for placebo. When survival in study
(any premature discontinuation, including for additional pharmacotherapy) was
evaluated, the difference between the treatment groups was significant (P ¼ 0.04).
For survival in study, the median survival times were 14 weeks for lamotrigine and
8 weeks for placebo. Time to additional pharmacotherapy and survival in study
did not yield significant differences between lamotrigine and placebo in patients
with BP-I disorder. When time to additional pharmacotherapy was evaluated, a
trend toward significance (P ¼ 0.07) was found in the separation between placebo
and lamotrigine. Median survival time without additional pharmacotherapy for
the BP-II subtype was 17 weeks for lamotrigine and 7 weeks for placebo. The
overall survival in study analysis yielded a significant separation between treat-
ment groups (P ¼ 0.01). Median overall survival was 15 weeks for lamotrigine and
4 weeks for placebo. The majority of those patients (80%) requiring additional
pharmacotherapy were treated for depressive symptoms; 20% were treated for
emerging manic, hypomanic, or mixed symptoms.
The percentage of patients who completed the 6-month randomized phase
clinically stable on monotherapy without evidence of relapse into hypomania,
mania, or depression was significantly greater in the lamotrigine group than in the
placebo group. Of the 60 patients who were stable for 6 months of monotherapy,
37 of 90 (41%) were in the lamotrigine group compared with 23 of 87 (26%) in the
placebo group (P ¼ 0.03). The difference for lamotrigine versus placebo was not
statistically significant for the BP-I subtype, but was significant (46% versus 18%,
73 Rapid-cycling bipolar disorder


respectively; P ¼ 0.04) for the bipolar II subtype. The CGI-S and GAS were used to
provide additional measures of clinical stability. For the overall study population
and the BP-I subtype, there were no statistically significant differences between
treatment groups in CGI-S change from baseline scores using the LOCF. For the
BP-II subtype, trends toward statistically significant differences (P < 0.10) favor-
ing the lamotrigine group were observed in CGI-S scores compared with the
placebo group at weeks 6 and 12. No statistically significant differences favoring
lamotrigine were observed between groups in GAS change from baseline scores in
the general cohort of patients (LOCF). Significant differences favoring lamotrigine
were noted at weeks 3, 6, and 12 in the BP-II subtype; however, no significant
differences were noted at any time point for the BP-I subtype. There were no
significant differences observed in the change from baseline LOCF analyses at any
point for the 17-item Hamilton-D or the Mania Rating Scale.
The most common adverse events (! 10%) observed during the open stabil-
ization phase were headache, infection, influenza, nausea, dream abnormality,
dizziness, and rash. In the randomized phase, 122 patients (67% lamotrigine; 68%

<<

. 13
( 68 .)



>>