Spectrum
of Efficacy
With
Joseph
of Valproate
Rapid-Cycling
R. Calabrese,
Bipolar
M.D.,
and
In order to explore vaiproate’s spectrum of efficacy in rapid-cycling bipolar disorder, 55 patients underwent a prospective, open, 7.8-month trial designed to assess the drug’s acute and prophylactic properties. Twenty patients received monotherapy, and 35 received combination therapy. Moderate to marked acute antidepressant responses were seen in 47% of the patients, prophylactic antidepressant responses in 76%, acute antimanic responses in 91 %, prophylactic antimanic responses in 94%, acute responses in mixed states in 85%, and prophylactic responses in mixed states in 93%. Consistent with other anticonvulsant literature, these data suggest that valproate has marked antimanic and mixed state efficacy, but minimal to moderate antidepressant properties. (Am J Psychiatry 1990; 147:431-434)
A
lthough lithium therapy continues to be viewed as the treatment of choice for bipolar affective disorder, 20%-40% of patients either do not tolerate this drug or do not respond to it. A growing body of data suggests that rapid-cycling bipolar disorder accounts for a significant proportion of bipolar affective disorder that is resistant to lithium. Among bipolar patients, 13%-20% experience rapid cycling (1, 2). Since frequency
of
recurrence
has
previously
been
believed
to
contribute substantially to morbidity (3), improving the clinical management of this refractory patient population has the potential of significantly reducing human suffering. Such anticonvulsants as carbamazepine are believed to have particular efficacy in the management of rapid-cycling bipolar disorder (4). Although it has been demonstrated that carbamazepine has marked acute antimanic properties, its acute antidepressant properties are less impressive, and its prophylactic efficacy has recently been called into question (5, 6). The clinical use of carbamazepine is, in addition, compromised by its ability to induce its own metabolism, as well as to induce the metabolism of a variety of
Received
June
Oct. 31, 1989. pitats
print
of
Cleveland,
requests
Hanna
J
1989;
revision
©
Psychiatry
received
the Department Case
Western
to Dr. Catabrese,
Pavilion,
Copyright
Am
14,
From
2040
Abington
1990
American
147:4,
Oct.
2, 1989;
of Psychiatry, Reserve
University.
Hos-
Address
University Hospitals of Cleveland, Rd., Cleveland, OH 44106. Psychiatric Association.
April
1990
re-
A. Delucchi,
different valproate)
drugs (i.e., by enhancing
system (7, 8). Although several as an
Patients
Disorder
Gustavo
alternative
M.D.
birth control the P-450 authors
pills, haloperidol, microsomal enzyme
have
suggested
to carbamazepine
in the
valproate management
of the patient with lithium-resistant rapid cycling, to our knowledge there have been no trials that have systematically studied large, homogeneous samples of patients with rapid-cycling bipolar disorder (9, 10). Since bipolar disorder is a recurrent illness accompanied by manic, depressed, and mixed states, the spectrum of efficacy in both the acute and prophylactic setting is of particular importance. Few studies have specifically evaluated the efficacy of valproate in these settings.
Although
intensive,
the
tern of complex,
this design
type may
response and multifaceted,
systematically
spectively
of drug more
trial
is more
accurately
long-term recurrent
explore
in the acute this refractory
labor-
reflect
pat-
outcome in illness. In order
valproate’s
spectrum
given
valproate
either
in
this to
of efficacy
and prophylactic settings, patients variant of bipolar disorder were
combination
with pro-
monotherapy
or
therapy.
METHOD After DSM-III-R trial of design. bipolar cluded a history and
giving
informed consent, 55 patients who met criteria for bipolar disorder were given a valproate in a prospective, longitudinal, open These patients were consecutively referred to a disorder specialty clinic. In order to be inin the study, each patient was required to have of classic rapid cycling, as defined by Dunner
Fieve
(1),
depressions,
i.e.,
a history
manias,
or
tent with DSM-III-R, hibited cycle frequencies were further described affective
instability
that the depression
symptom were
nod.
accepted
University
in 55
Before
function
frequency, with
life
interviewing
entry,
lability)
of
screening
major Consis-
were
and
severity
were
techniques
and
member.
such
exand of
severity
baseline
SMA-16
a family
more year.
both hypomania the same 24-hour
and
duration, charting
or
patients with mixed states too numerous to count as having discrete episodes (mood
CBC,
four
per
complex of present within
study
testing,
of
hypomanias
were
Patients
and pethyroid
done.
Cycle
ascertained confirmed
were
by
excluded
431
EFFICACY
OF
VALPROATE
TABLE 1. Characteristics lar Disorder
IN
BIPOLAR
of 55 Patients
DISORDER
With Rapid-Cycling
Bipo-
to valproate in acute and prophylactic settings, he or she would be likely to remain in complete remission. Since the prophylactic antimanic effects of a drug do not equal its acute effects, we cannot assume that successful prophylaxis against mania is equivalent to
Item
N
%
Women Men
32 23
58.1 41.8
38
69.1
Self-referred
17
30.9
Bipolar type I Bipolar type II Mixed states Previously hospitalized
25 30 14 38
45.5 54.5 25.5 69.1
sets were Conversely,
Unemployed due to illness Overt thyroid failure
22 4
40.0 7.3
prophylactic
41
74.5
35
14
63.6 25.5
17
30.9
49 27
89.1 49.1
Physician-referred, treatment failures
Family
history
bance Treatment Lithium
previous
of mood
distur-
failures alone
Carbamazepine
alone
Lithiumlcarbamazepine combination
Current Current
severe severe
depressions manias
Characteristics increasing illness onset Number of cycles per Depression severity Mania severity Duration of depression Duration of mania were
apatients
as mild,
asked
year
46 32 26 22 16 the
severity
of their
83.6 58.2 47.3 40.0 29.1 current
mood
states
or severe.
but with persistence mild improvement;
proate responders exhibited moderate study
assessed
longitudinal
were defined or marked
spectrum settings,
of breakthrough and no response.
of
as those responses.
efficacy
denominators
patients Because
in both describing
acute
mood Va!-
who this
432
the
This
acute sample
larger than those some patients
acute
was
effects
patient
would
antimanic sizes for for who
were
never
properties prophylactic
acute failed
be
of valdata
efficacy data sets. in a trial of val-
to participate in the maintenance therapy could not be justified when acute responses were negligible. Anecdotal experience suggests that this chronically ill patient population commonly experiences substantial degrees of learned helplessness. As a result, interepisode psychosocial impairment remains after a complete response. For this reason, psychosocial function was not used as an outcome measure. A regimen of divalproex sodium enteric-coated tab-
unable
assessment;
started
i.e.
at 250
mg
,
b.i.d.
and
increased
by
250
mg every 5-7 days, as tolerated. Valproate total levels, CBC, and repeat SMA-16 were obtained at baseline, 1 month, and 3 months during the first 6 months, and every 9-12 months thereafter. When valproate and carbamazepine were used together, free levels of each were monitored, since valproate is highly protein bound noted
to
displace
protein-bound
carbamazepine.
In addition to receiving pharmacotherapy, tients participated in a wide range of interventions designed to obtain optimum and
assist
tive
psychotherapy,
in rehabilitation.
outpatient and
pressive
and
involvement
Manic
Depressive
in vivo
interventions
rehabilita-
therapy,
specialized
for
in
patients
the
used
and
National
Association.
desensitization,
were
study panonbiological compliance
included
groups
local
Anonymous,
ioral
These
marital/family
manic-depressive
families,
De-
Alcoholics
and
other
behav-
as indicated.
RESULTS Thirty-five resistant
of 55 patients to
patients
either
(36.4%)
(63.6%)
lithium
or
received
were
noted
carbamazepine.
valproate
to be Twenty
monotherapy
and
combination therapy. Of the 35 who received combination therapy, 21 (60.0%) received valproate augmentation of lithium, five (14.3%) received unimodal antidepressants (e.g. , tricyclic antidepressants, monoamine oxidase inhibitors), three (8.6%) valproate augmentation of lithium/carbamazepine 35
(63.6%)
therapy,
and
carbamazepine. ing drug trial
and
dian=5.5
out-
for
come vary, since information concerning acute efficacy was obtainable only for those patients who presented with symptoms of the mood state in question. For example, if a patient presented depressed and responded
efficacy.
to test Therefore,
proate’s
and
from the study if they presented with elevated liver function test results or evidence of other CNS abnormalities. Valproate had never been previously prescribed for these patients. Previous bimodal therapy (lithium and carbamazepine) was classified as a failure if patients had experienced a major episode of depression, mania, or hypomania meeting DSM-III-R criteria or intolerable side effects. Table 1 provides additional information for the total sample. The mean±SD age of the patients at entry to the study was 40.7± 13.4 years; the mean age at onset of illness was 19.3±11.2 years. The mean number of cycles per year was 9.2±9.7 (median=6). The mean durations of the patients’ current depressions and manias were 46.0±68.0 days and 29.9±61.3 days, respectively. Outcome measures were defined as follows: marked improvement-complete cessation of all cycling (i.e., no breakthroughs of any kind); moderate improvement-improvement in cycle severity, duration, or frequency swings;
antimanic
available proate.
lets
since
to rate
moderate,
acute
was
an
one
(2.9%)
months). average
valproate
augmentation
of
The mean±SD duration of this was 7.8±5.9 months (range=1-24, Patients of 21.4
1686±680
.tg/ml. Table
2 describes
for both
the entire
mg;
the analyses
J
had
The
mean
sample
Am
had
years.
evidence
mean
serum based
on
of illness
valproate level
was
outcome
147:4,
dose 84±28 criteria
of valproate-treated
Psychiatry
ongome-
patients
April
1990
JOSEPH
TABLE 2. Outcome of Patients With Rapid-Cycling Bipolar der Who Received Vaiproate With or Without Other Drugs
Disor-
Responders Marked Response
Mood State and Type of Treatment
Total
N
%
N
who received (N=20)
Depression Acute
Prophylactic Mixed states Acute Prophylactic on
number
34 Si
31 48
91.2 94.1
19 40
61.3 83.3
34 51
16 39
47.1 76.5
13
81.3
21
53.8
14 14
12 13
85.7 92.9
12 13
100.0 100.0
10 20
9 19
90.0 95.0
8 17
88.9 89.5
11
20
7 17
63.6 85.0
4 11
57.1 64.7
6 6
6 6
valproate
Mania Acute Prophylactic
aBased
100.0 100.0
6 6
100.0 100.0
of responders.
and the 20 patients who received monotherapy. All three of the patients who did not respond to the acute antimanic properties of valproate had bipolar I rapid cycling. Of the patients with type I and II disorders, 84% (N=21) versus 93.3% (N28) were responders in the acute setting; and 55.8% (N=14) versus 77.0% (N=23) exhibited marked responses. Comorbidity was prevalent, with 15 of 55 (27.3%) exhibiting statedependent alcohol abuse, primary alcoholism in four (7.3%), other substance abuse in 12 (21.8%), panic attacks in 22 (40.0%), generalized anxiety in nine (16.4%), and obsessive-compulsive disorder in one (1.8%). Panic attacks subsided in 21 of 22 patients. State-dependent alcohol abuse remitted in 15 of 15 patients. Other substance abuse remitted in 12 of 12 patients. Dependence on alcohol unrelated to mood disturbance was treated with Alcoholics Anonymous in each of four patients; one of the four did not maintam sobriety. Non-state-dependent generalized anxiety remitted in eight of nine patients. Thyroid supplementation was continued in the four patients who had reported a previously treated hypothyroidism. Twenty-nine of the 55 patients (52.7%) had side effects. There were eight reports of nausea, six of epigastric cramping, five of tremors, four of ataxia, four of lethargy, three of alopecia that was dose-related and resolved once the dose was decreased, and one each of slurred speech, blurred vision, rash, headache, enuresis, fluid retention, and hepatotoxicity. The drug was discontinued twice due to gastrointestinal upset, once due to alopecia, once due to de novo elevations of liver function test results, accompanied by signs of liver injury, in a patient
Am
J
Psychiatry
GUSTAVO
A. DELUCCHI
with acquired immune deficiency syndrome, and five times due to lack of antidepressant efficacy. The latter involved partial responders to lithium who had already experienced complete remission of their manias but were given valproate to treat residual major depression.
DISCUSSION
group (NSS) Mania Acute Prophylactic Depression Acute Prophylactic Mixed states Acute Prophylactic atone
AND
%
Total
Patients
R. CALABRESE
147:4,
April
1990
Until recently, inadequate attention has been given to the spectrum of drug efficacy in the treatment of bipolar disorder. As the rapid-cycling variant of bipolar disorder becomes increasingly well-recognized, the spectrum of efficacy in both the acute and longitudinal settings will take on increasing importance. It is of particular concern to patients with the rapid-cycling variant. Before they are willing to relinquish their highs through effective antimanic therapy, many will query the physician as to the drug’s antidepressant efficacy. They realize the currently available bimodal therapies are better antimanic agents than antidepressants. A unimodal assessment of an agent’s antimanic properties in a placebo-controlled, double-blind setting is, therefore, complemented by longitudinal, naturalistic studies that individually assess manic, mixed, and depressed mood states in both the acute and prophylactic settings. Unfortunately, longitudinal, doubleblind, placebo-controlled drug trials of this kind are expensive, critically reviewed by institutional review boards, and labor-intensive. Large sample sizes are required to accumulate adequate acute cohorts for the individual mood states, and study periods are lengthy. This prospective, open, longitudinal study of valproate monotherapy and combination therapy was designed to address some of these concerns. Consistent with other anticonvulsant literature, our data suggest that valproate possesses potent antimanic and mixed state properties in both the acute and prophylactic settings but minimal to moderate antidepressant properties. The frequency of prior psychiatric hospitalizations, resistance to lithium/carbamazepine treatment, bipolar type I disease, and unemployment suggest that these patients were quite ill and did not have merely less serious variants of mood disturbance, such as cyclothymia. Patients’ reports with respect to years since onset of illness (21.4) and duration and severity of illness also suggested significant levels of impairment (see table 1). Of particular interest is the observation that 84% of the patients reported that their disease became more autonomous as it “matured.” They noted that the cycles became more frequent and had no apparent relationship to any stressful life event. This feature, in particular, has been viewed in the past as a predictor of poor outcome. Approximately half viewed the amplitude of their manias and depressions as increasing. Nearly half reported increased durations of depressions over the course of their illness, and over one-fourth reported increased durations of manias. Due to the ambiguity of our DSM-III-R criteria for
433
EFFICACY
OF VALPROATE
IN BIPOLAR
DISORDER
mixed states, the data concerning valproate’s marked efficacy in mixed states in our sample of 14 patients must be viewed with caution. As the mean duration of these mixed states was 18.7±23.9 days (range= 1-70 days), these episodes were not just short transitional states as frequently seen when patients cycle from a high
to
a low
or
vice
versa.
These
subjects
cycled
tively
434
that
to this
can
be
outpatient
given
both
population
serially
and
and
used
REFERENCES
into
mixed states that persisted for substantial periods of time and exhibited the best outcome. In fact, some would specify that these patients be more accurately referred to as ultrafast rapid cyclers, i.e., having cycles of less than 48 hours (1 1). Since in most cases their cycle frequencies were too numerous to count, we elected to stay within the DSM-III-R criteria. Our findings for this mood state, however, are consistent with the literature, which suggests that anticonvulsants are particularly helpful in the management of mixed states (12). Of particular note is valproate’s marked antimanic efficacy in this otherwise refractory patient population. Although the antidepressant properties of this drug were clear-cut, they were less impressive and again suggest that we have yet another bimodal agent with minimal to moderate antidepressant efficacy. For this reason, the challenge of treating the depressed phase of rapid-cycling bipolar disorder persists. Should conventional unimodal antidepressants be used and risk exacerbating cycle frequency (13), or should bimodal therapies (i.e., lithium, carbamazepine, valproate) be relied on solely? Future drug trials might be directed toward the development of well-tolerated bimodal pharmacotherapies that possess potent antidepressant properties. In addition, a self-administered rating scale is needed
tify cycle frequency, severity, and duration. Such an instrument is currently under development and will assist in the systematic evaluation of the periodic phenomena of rapid-cycling in affective disorders.
1 . Dunner 2.
failure. Dunner
DL, Fieve Arch DL,
depressive 3. Kukoputos
RR:
Clinical
factors
in lithium
prophylaxis
Gen Psychiatry 1974; 30:229-233 Patrick D, Stattone F, et at: Rapid-cycling
patients. Compr Psychiatry A, Reginaldi D, Laddomada
manic-depressive
cycle
and
changes
1977; 18:561-565 P, et al: Course caused
by treatment.
manic
of the Phar-
macopsychiatry 1980; 13:156-167 4. Post RM, Uhde TW, Ballenger JC, et at: Prophylactic efficacy of carbamazepine in manic-depressive illness. Am J Psychiatry 1983; 140:1602-1604 S. Frankenburg FR, Tohen M, Cohen BM, et at: Long-term response to carbamazepine: a retrospective study. J CIin Psychopharmacol 1988; 8:130-132 6. Pnien RF, Getenberg AJ: Alternatives to lithium for preventive treatment of bipolar disorder. Am J Psychiatry 1989; 146:840848
7. Mikati MA, Browne pine autoinduction. 8.
Hansten
PD,
Horn
TR, Collins JF: Time course Neurology 1989; 39:592-594 JR:
Inhibition
of oral
of carbamaze-
contraceptive
efficacy.
Drug Interactions Newsletter 1985; 5:7-10 9. McElroy SL, Keck PE, Pope HG, et at: Vatproate in the treatment of rapid-cycling bipolar disorder. J Clin Psychopharmacol 1988; 8:275-279 10. Calabrese JR, Delucchi G: Phenomenotogy of rapid-cycling manic depression and its treatment with valproate. J Clin Psychiatry 1989; 50:30-34 1 1. Alarcon R: Rapid-cycling affective disorders: a clinical review. Compr
12. 13.
Psychiatry
1985;
26:522-540
Himmelhoch JM, Garfinkel ME: Mixed mania: diagnosis and treatment. Psychopharmacol Butt 1986; 22:613-620 Wehr TA, Goodwin FK: Rapid cycling in manic depressives
prospec-
induced
to quan-
36:555-559
by tnicyclic
antidepressants.
Am
J
Psychiatry
Arch
Gen
147:4,
Psychiatry
1979;
April
1990