THANK YOU FOR YOUR ORDER

Thank you for your recent purchase. If you need further assistance, please contact Customer Service at
abbottsupport@copyright.com. Please include the Order ID, so we can better assist you.

This is not an invoice.

CUSTOMER INFORMATION ORDER INFORMATION

Ordered For: Thais dos Anjos Infotrieve Order ID: 14140721
Company: Ordered For: Thais dos Anjos
Client ID: Ordered For Email: thais.dosanjos@abbott.com
Address: Ordered: 10/7/2020 12:17 PM
Deliver Via: Email (PDF)

Sao Paulo, Sao Paulo 04566-905 Delivery Address: thais.dosanjos@abbott.com

Country: Brazil Ledger: 118067-BR-EPD-MEDICAL MANAGEMENT

Phone: Department: 118067
Fax: Notes for Myself:
Email: thais.dosanjos@abbott.com

DOCUMENT INFORMATION
Std. Num.: 00987484 Type: Doc Del (Journal Article)
Publication: Copies:
JAMA : The Journal of the American Medical Association 1
Publisher: American Medical Association Urgency: Normal
Vol(Iss) Pg: 271 (12) p.918-24 Genre: Journal Article
Date 3/1994 Total Fee: $40.50

Title: Efficacy of divalproex vs lithium and placebo in the treatment of mania. The Depakote Mania Study Group

Author(s): Bowden, C L; Brugger, A M; Swann, A C; Calabrese, J R; Janicak, P G; Petty, F; Dilsaver, S C; Davis, J M; Rush, A J;
Small, J G

Usage: Ordering 1 copy for the following stated purpose: "Use a copy for myself". Covers viewing or saving content locally on
Abbott-owned devices.

The contents of the attached document are copyrighted works. You have secured permission to use this document for the following
purpose: Ordering 1 copy for the following stated purpose: "Use a copy for myself"

You have not secured permission through Copyright Clearance Center, Inc. for any other purpose but may have other rights pursuant to
other arrangements you may have with the copyright owner or an authorized licensing body. To the extent that a publisher or other
appropriate rights-holder has placed additional terms and conditions on your use of this document, such terms and conditions are
specified herein under “Copyright Terms”. If you need to secure additional permission with respect to this content, please
purchase the appropriate permission via RightFind.
Copyright Terms:

(800) 422-4633 Toll Free Email: abbottsupport@copyright.com

(203) 423-2175 Direct
 Lithium and
Efficacy of Divalproex vs
Placebo in the Treatment of Mania
Charles L. Bowden, MD; Andrew M. Brugger, MD; Alan C. Swann, MD; Joseph R. Calabrese, MD;
Philip G. Janicak, MD; Frederick Petty, MD, PhD; Steven C. Dilsaver, MD; John M. Davis, MD;
A. John Rush, MD; Joyce G. Small, MD; Enrique S. Garza-Trevi\l=n~\o,MD; S. Craig Risch, MD;
Paul J. Goodnick, MD; David D. Morris, PhD; for the Depakote Mania Study Group

Objective.\p=m-\Tocompare the effectiveness of divalproex sodium with that of MANIC-depressive illness (bipolar dis¬
lithium and placebo in patients with acute mania. order) is a chronic, cyclic disease that
afflicts approximately 1% of the popu¬
Design.\p=m-\Randomized,double-blind, parallel-group study of treatment out- lation.13 The manic episodes are par¬
comes in patients with manic-depressive illness.
Patients.\p=m-\Atotal of 179 hospitalized, acutely manic patients meeting the Re- ticularly disruptive, with symptoms such
as hyperactivity, explosive temper, im¬
search Diagnostic Criteria for manic disorder, approximately half of whom had been paired judgment, insomnia, disorganized
nonresponsive to lithium previously, were studied at nine university-affiliated hos- behavior, hypersexuality, grandiosity,
pitals. and often delusions, leading to severe
Interventions.\p=m-\Aftera minimum 3-day washout period, random assignment for functional impairment as manifested by
21 days to divalproex, lithium, or placebo in a 2:1:2 ratio. Dosage of divalproex and alienation from family, friends, and co-
lithium was increased if tolerated to a target concentration of 1041 \g=m\mol/L (150 \g=m\g/ workers; indebtedness; job loss; divorce;
mL) or 1.5 mmol/L (conventionally expressed as milliequivalents per liter), respec- and other problems of living.4
The predisposition to this mood dis¬
tively. order is usually genetically determined,
Main Outcome Measures.\p=m-\Primaryoutcome measures were changes in the
Mania Rating scale derived from the Schedule for Affective Disorders and Schizo- although persons may develop forms of
bipolar disorder secondary to various
phrenia. medical conditions, including the se¬
Results.\p=m-\Intent-to-treatanalysis for efficacy was based on data from 68, 35, and
quelae of head trauma.5·6 Manic-depres¬
73 patients in the divalproex, lithium, and placebo groups, respectively. Groups sive illness frequently begins in the teen
were initially comparable except that all eight patients with four or more manic epi- years but often escapes diagnosis at this
sodes in the previous year were in the divalproex group. In 30%, 33%, and 51% time because episodes are misinter¬
of the above groups, treatment was prematurely terminated due to lack of efficacy, preted as conduct disorder, schizophre¬
with fewer premature terminations from divalproex than placebo (P=.017). The nia, depression, or other disorders.7,8
From 35% to 50% of bipolar patients
proportions of patients improving at least 50% were higher for divalproex and lithium abuse alcohol, are alcoholics, or abuse
groups than for the placebo group: 48% for divalproex (P=.004) and 49% for lithium other drugs.1·911 Suicide mortality is
(P=.025) vs 25% for placebo. Divalproex was as effective in rapid-cycling manic about 15% in untreated persons with
patients as in other patients. manic-depressive illness. It is estimated
Conclusions.\p=m-\Bothdivalproex and lithium were significantly more effective that only one in three persons with the
than placebo in reducing the symptoms of acute mania. The efficacy of divalproex disorder receives treatment.12
appears to be independent of prior responsiveness to lithium. Lithium, the only drug approved by
(JAMA. 1994;271:918-924) the US Food and Drug Administration
for the treatment of acute mania, is in¬
effective or poorly tolerated in at least
From the Department of Psychiatry, University of Other members of the Depakote Mania Study Group
one third oftreated patients and has the
Texas Health Science Center and Audie L. Murphy are listed at the end of this article. Drs Brugger and
Memorial Veterans Hospital, San Antonio (Drs Bowden Morris purchase Abbott Laboratories stock on a regu- narrowest gap between therapeutic and
and Garza-Trevi\l=n~\o);Clinical Research Section, Abbott lar basis as part of their retirement plans; Drs Bowden, toxic concentrations of any agent rou¬
Laboratories, North Chicago, Ill (Drs Brugger and Mor- Swann, Calabrese, and Janicak have received hono-
ris); Department of Psychiatry, University of Texas raria from Abbott for educational programs; Dr Janicak tinely prescribed in psychiatric prac¬
Medical School, Houston (Drs Swann and Dilsaver); owns equities in Abbott Laboratories through an in- tice.13"18 Determining the efficacy of
Department of Psychiatry, Case Western Reserve Uni- vestment plan for retirement funds. Drs Davis and lithium in the acute manic phase of bi¬
versity School of Medicine, Cleveland, Ohio (Dr Cala- Goodnick have received remuneration from Abbott as
brese); Illinois State Psychiatric Institute, Chicago, and speakers. Dr Small owns stock in Abbott Laboratories polar disorder is problematic, as these
and has received honoraria for participation in educa- patients are difficult to study and pla¬
Department of Psychiatry, University of Illinois at Chi-
cago (Drs Janicak and Davis); Veterans Affairs Medi- tional programs. Dr Garza-Trevi\l=n~\ohas received remu- cebo controls are impractical for all but
cal Center and Department of Psychiatry, University of neration from Abbott for participation in continuing
Texas Southwestern Medical Center, Dallas (Drs Petty medical education activities. Dr Risch has received re- short periods due to the severity of the
and Rush); Department of Psychiatry, Indiana Univer- muneration from Abbott as a speaker and a consultant. illness. The most recent placebo-con¬
sity School of Medicine, Indianapolis (Dr Small); De- Reprint requests to Division of Biological Psychiatry, trolled study of lithium in manic patients
partment of Psychiatry, Emory University School of University of Texas Health Science Center at San An- was reported more than two decades
Medicine, Atlanta, Ga (Dr Risch); and Department of tonio, 7703 Floyd Curl Dr, San Antonio, TX 78284-7792
Psychiatry, University of Miami (Fla) (Dr Goodnick). (Dr Bowden). ago,19 and there have been no placebo-

Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020

 controlled studies with a parallel-group
design.20·21
In the 1970s, researchers began to
investigate anticonvulsant drugs in the
treatment of manic states, particularly
carbamazepine, clonazepam, and val-
proate.22·23 Divalproex sodium was used
in the present study, as it had been found
to be effective in previous controlled
trials of manic patients.24·25
Herein we report the first random¬
ized, double-blind, parallel-group inves¬
tigation of divalproex with a lithium arm
and a placebo control. In addition to as¬
sessing the efficacy of divalproex vs
lithium and placebo, we address several
other questions: (1) Which components
of manic symptomatology are most re¬
sponsive to divalproex, lithium, or both?
(2) Does the adverse effect profile in
divalproex-treated patients differ from
that in lithium-treated patients? (3) Are
differences in response related to the
presence of rapid cycling or prior esti¬
mated response to lithium?
METHODS
Patients
The patients we studied were hospi¬
talized at the following institutions: Au-
die L. Murphy Memorial Veterans Hos¬
pital, San Antonio, Tex, and San Anto¬
nio State Hospital (44 patients); Harris
County Psychiatric Center, Houston,
Tex (49 patients); Larue D. Carter Me¬
morial Hospital, Indianapolis, Ind (19
patients); Emory University Hospital,
Atlanta, Ga (six patients); University
Hospitals of Cleveland (Ohio) (22 pa¬
tients); Veterans Affairs Medical Cen¬
ter, Dallas, Tex (13 patients); Jackson
Memorial Hospital, Miami, Fla (seven
patients); and Illinois State Psychiatric
Center, Chicago (16 patients). Three pa¬
tients at an additional site were included
in the safety analysis but were excluded
from the intent-to-treat data set because
the center enrolled only those three pa¬
tients, a number insufficient to meet the
requirements for the van Eiteren sta¬
tistic used in the primary analysis.26 En¬
rolled in the study were 179 men and
women between 18 and 65 years of age
who met Research Diagnostic Criteria
for manic disorder, based on the struc¬
tured interview and rating scale of the
Schedule for Affective Disorders and
Schizophrenia (SADS).27 Patients also
had Mania Rating scale scores of at least
14 on the last washout day with scores
of 2 or higher on at least four items and
had undetectable serum lithium concen¬
trations prior to randomization. Scores
on the SADS and SADS-C (SADS-
Change version) were adjusted so that
the absence of a symptom was scored as
0 instead of 1, to facilitate the interpre-
Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020
tation of change scores. Change scores
were calculated by subtracting the score
on a variable at the final evaluation from
the baseline score, defined as the last
evaluation performed prior to adminis¬
tration of the first dose of blinded study
drug.
Reasons for exclusion were as follows:
(1) a history of severe side effects from
lithium treatment, (2) prior treatment
with divalproex or valproic acid, (3)
schneiderian first-rank symptoms occur¬
ring throughout the day without manic
symptoms for several days or intermit¬
tently for more than a week, (4) central
nervous system or neuromuscular dis¬
orders, (5) uncontrolled diseases, (6)
drug-induced mania or mania induced
by the acquired immunodeficiency syn¬
drome, (7) substance abuse as defined
by the SADS, (8) a positive result on
any toxicology screening tests, (9) con¬
comitant treatment with medications
that would confound the results, and
(10) pregnancy.
The study objectives were explained
to the patients and each gave signed
informed consent. The study was ap¬
proved by the institutional review board
of each center.
Patients were terminated from the
study if they required psychoactive
medications not specified in the proto¬
col, developed severe depression, or had
an increase of greater than 30% in the
SADS-C Mania Rating scale over that
measured on the last day of the washout
period. They could also have been ter¬
minated if they met all of the following
three recovery criteria: (1) a reduction
of at least 50% from baseline (ie, the last
day of the washout period) in their
SADS-C Mania Rating scale score; (2)
no SADS-C Mania Rating scale item
score higher than 2; and (3) a Global
Assessment Scale (GAS) score higher
than 70. The poststudy treatment was
individually determined for each patient
without breaking the study blind.
Study Design
This randomized, multicenter, double-
blind, parallel-group study of hospital¬
ized patients in the manic phase of bi¬
polar disorder was designed to evaluate
the efficacy and safety of divalproex so¬
dium (Depakote, Abbott Laboratories,
North Chicago, 111), an oligomeric com¬
plex of sodium valproate and valproic
acid in a 1:1 molar ratio. Comparison
and control groups consisted of patients
treated with lithium carbonate (Lith-
ane, Miles Pharmaceuticals, West Ha¬
ven, Conn) and placebo.
Patients were admitted to the hospi¬
tal on (or before) day 1 of an initial 3- to
21-day washout period. The washout pe¬
riod consisted of the longer of 3 days or
five half-lives of the psychoactive drug
taken on admission with the longest half-
life. For patients who had previously
taken lithium, the prior efficacy and tol-
erability of the drug were assessed dur¬
ing a patient interview and, when avail¬
able, from medical records.
Screening tests during the washout
period included a complete medical and
psychiatric history; physical examina¬
tion; standard laboratory tests, includ¬
ing assessment of thyroid and liver func¬
tion; and, in patients more than 40 years
of age, an electrocardiogram. A toxicol¬
ogy screen for cocaine, phencyclidine,
and amphetamines was performed.
When applicable, a pregnancy test was
performed and was repeated during
treatment. On the last study day, the
physical examination and laboratory
evaluations, including measurement of
serum valproate and lithium concentra¬
tions, were repeated. Concurrent medi¬
cations were recorded daily.
On completion of the washout period,
patients who still exhibited a manic epi¬
sode as defined by protocol-specified cri¬
teria qualified for entry into the 21-day
experimental period. Patients were ran¬
domly assigned to divalproex, lithium,
or placebo in a 2:1:2 ratio. A separate
randomization schedule for each center
was generated prior to the study start.
Centers were sent patient numbers in
blocks of 10; unknown to the investiga¬
tors, treatment group assignments were
randomized in blocks of five within each
set of 10 numbers. Divalproex sodium
was administered at an initial dose of
750 mg/d and lithium carbonate at 900
mg/d. Study medication (active drug or
placebo) was dispensed in divided doses
three times daily, 30 minutes after meals,
in identical-appearing capsules. On day
3, the total daily dosages of divalproex
or lithium were increased to 1000 mg
and 1200 mg, respectively, and trough
serum concentrations ofboth drugs were
determined. On day 5, an unblinded phy¬
sician at each center reviewed the se¬
rum concentration and adjusted the dos¬
age of active medication. Thereafter,
trough concentrations were measured
on days 8,10,12,15, and 18. Medication
adjustments were made on days 7, 10,
12, 14, and 17. Drug dosage was raised
on each adjustment day unless precluded
by an adverse event or a serum concen¬
tration of valproate or lithium exceed¬
ing 1041 µ /L (150 µg/mL) or 1.5
mmol/L (conventionally expressed as
milliequivalents per liter), respectively.
Comparable adjustments were made in
the dosage of placebo according to
blinded protocol-specified dosing sched¬
ules.
The protocol allowed the use of ad-
junctive chloral hydrate or lorazepam
 Table 1.—Characteristics of the Three
Characteristic
ge, y (mean±SD)
Sex, % male:female
Race, % whlte:black:other
Length of illness, y (mean ± SD)
Bipolar episodes occurring
a4/y in past 2 y, No. (%)
Any major mood episode*
Mania only*
Prior lithium treatment, No. (%)
Effective, toleratedt
Effective, not toleratedt
Ineffective, toleratedt
Ineffective, not toleratedt
*Percentages are based on total number of patients with available histories in each group (n=57, 57, and 26 in
placebo, divalproex, and lithium groups, respectively).
tPatients in the divalproex group had a significantly higher incidence of manic episodes than patients In the lithium
or placebo groups (P<.05).
^Percentages are based on numbers of patients who took lithium previously.
as needed for control of agitation, irri¬
tability, restlessness, insomnia, and
hostile behaviors. The maximum daily
dosages of chloral hydrate and loraze-
pam were 4 g and 2 mg, respectively,
through treatment day 4, then 2 g and
1 mg, respectively, through day 10.
These medications were not permitted
during the 8 hours before behavioral
assessments. Neuroleptic drugs were
not allowed in the protocol.
Psychiatric Rating Scales and
Behavioral Assessment
On day 1 of the washout period, the
manic syndrome and overall psychiatric
symptomatology were evaluated by the
SADS, and the Research Diagnostic Cri¬
teria diagnosis was made. On the last
day of washout, the psychiatrist rated
the patient using the SADS-C and the
GAS. Whenever possible, the same
blinded investigator rated the patient
throughout the study. The psychiatric
nurse administered the Affective Dis¬
order Rating Scale (ADRS) on days
3,5,7,9,12,18, and 20. During the treat¬
ment period, the SADS-C and GAS were
administered on days 5, 10, 15, and 21.
The psychiatrists and nurses who
were to rate patients participated in
training sessions in which videotapes of
patient assessments were shown. Be¬
fore they could serve as study raters,
rating investigators had to achieve an
intraclass correlation of at least 0.50 with
the original set of raters on the SADS
mania items; most scored 0.70 or higher.
(A correlation between 0.40 and 0.74 rep¬
resents fair to good agreement.2*) An
additional assessment of interrater re¬
liability was made midway through the
study. Correlations for the manic items
ranged from 0.86 to 0.91 within centers
and 0.89 across centers. In addition to
Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020
Groups of Patients Before Treatment
Treatment Group
Divalproex Lithium
(n=69) (n=36)
39.1 ±11.2
52:48
66:17:17
18.0±12.4 16.1 ±11.0
11(19) 1(4)
8(14)t 0 tion were assessed with Fisher's Exact
54 (78) 31 (86)
22(41) 16(52)
7(13)
11
19(35) (35)
6(11) 4(13)
videotape training, printed guidelines
for the use of the rating scales were
furnished to all raters.
The Mania Rating scale score was the
primary efficacy variable. This scale com¬
prises two subscales based on SADS
items plus an additional question on im¬
paired insight; possible scores range
from 0 to 52. The Manic Syndrome sub-
scale score is the total of five items—
elevated mood, less need for sleep, ex¬
cessive energy, excessive activity, and
grandiosity; possible scores range from
0 to 25. The Behavior and Ideation sub-
scale score is the total of five items—
irritability, motor hyperactivity, accel¬
erated speech, racing thoughts, and poor
judgment; possible scores range from 0
to 22. No cutoffs have been established
for "normality."
The SADS items constituting the Ma¬
nia Rating scale were selected as a pri¬
mary rating measure because of their
high interrater and test-retest reliabil¬
ity,29 ability to distinguish manic patients
from nonmanic subjects and healthy con¬
trols,30 and sensitivity to change with
pharmacotherapy.31 The Mania Rating
scale was scored by the psychiatrist on
the basis of all available clinical data.
Supportive efficacy variables included
the GAS score; individual items of the
Mania Rating scale (listed above); and
ADRS factors for mania, elation, para-
noia/destructiveness, depression, psy-
chomotor retardation, anxiety/agitation,
anger, psychosis, and confusion. The
ADRS was scored on the basis of nurses'
observations of patient behavior.
Statistical Analyses
All tests were two-tailed and analy¬
ses were performed using SAS soft¬
ware.32 Comparability of baseline data
among treatment groups (demographic
data, psychiatric history) was assessed
for quantitative variables by the overall
F test of one-way analysis of variance
Placebo
(n=74)
(ANOVA) and for qualitative variables
by Fisher's Exact Test. Length of bi¬
57:43 polar illness, as defined by number of
72:19:9 years from onset of first major mood
episode (maniaor majordepression), was
assessed with the Kruskal-Wallis test.
Pairwise treatment group differences
6(10) for each reason for premature termina¬
61 (82) Test extended to r x c tables.33 Pairwise
19(31) treatment group differences for changes
in efficacy variables from baseline to
6(10)
31 (51)
final evaluation were assessed by the
van Eiteren method of linearly combin¬
5(8)
ing Wilcoxon test results from indepen¬
dent strata (ie, investigators).2'1 A sepa¬
rate van Eiteren analysis, based on pa¬
tients assigned to relevant treatment
groups, was performed for each pair-
wise comparison (divalproex vs placebo,
lithium vs placebo, divalproex vs
lithium). The Cochran-Mantel-Haenszel
test was used to assess pairwise treat¬
ment group differences for all Mania Rat¬
ing scale items, with investigators form¬
ing the independent strata. This test
was also used to evaluate pairwise treat¬
ment group differences in the propor¬
tions of patients achieving a 50% im¬
provement from baseline. The van
Eiteren test was used in the carry-for¬
ward analysis of change from baseline
to days 5, 10, 15, and 21. Treatment-
emergent adverse events, categorized
by COSTART terms,34 were assessed
for pairwise treatment group differences
in the proportion of patients reporting
each event. A treatment-emergent ad¬
verse event was defined as any event
beginning on or after the first dose of
blinded study drug. Changes in labora¬
tory findings were assessed for pair-
wise treatment group differences using
the one-way ANOVA, using the least
significant difference method. There
were no significant center-by-treatment-
group interactions for the primary ef¬
ficacy variables.
RESULTS
Baseline Characteristics
A total of 179 patients were random¬
ized to treatment groups: 69 to dival¬
proex, 36 to lithium, and 74 to placebo.
Treatment groups did not differ signifi¬
cantly in demographic characteristics.
Medical histories were similar (Table 1),
except that all of the eight patients who
averaged four or more manic episodes
per year for the past 2 years fell into the
divalproex group. The efficacy results
were essentially unchanged when these
patients were removed from the analy¬
ses. Four (50%) of these eight patients
 Table 2.—Reasons for Premature Termination

Treatment Group, No. (%)

-1
Divalproex Lithium Placebo
Reason (n=69) (n=36) (n=74) Lithium •M· · mX·* t· t M I M t· I
Lack of efficacy 21 (30)* 12(33) 38(51)

.^j^
Intolerance to
treatment 4(6) 4(11) 2(3)
Met recovery
criteria 3(4) 2(6) 2(3) Divalproex ...
Noncompliance KD 1(3) 3(4)
Intercurrent illness
Administrative
0 1(3) 0 1
reason 4(6) 2(6) 2(3) Placebo
Total 33 (48)t 22(61) 47 (64)

*Proportion of terminations was significantly lower -1-1-1-1—

than in the placebo group (P=.017).
tTotal percentages were rounded to the nearest whole +20 10 0 -10 -20 -30 -40
number. in Score From Baseline
Change

Fig 2.—Scattergram showing changes in individual scores (each circle representing one patient) from
28 baseline to final evaluation
on Mania Rating scale. Arrows indicate median score change.

26
24 -O
22
20
18-
16
^-!-1- -
Baseline 5 10 15 21
Time on Protocol, d
Fig 1.—Changes from baseline to final evaluation In
Mania Rating scale score, Schedule for Affective
Disorders and Schizophrenia-Change version
(SADS-C). Solid squares indicate divalproex; solid
circles, lithium; and open circles, placebo. Numbers
on the vertical axis indicate the sum of all Items on
this subscale of the SADS-C. Asterisks and dagger
indicate time points at which a significant difference
(P<.05) was observed between divalproex and
lithium, respectively, and placebo treatment.
had at least 50% improvement on the
Manic Syndrome subscale score, which
was comparable to the response rate
(48%) seen for the divalproex group as
a whole. Otherwise, treatment groups
were comparable: 89% had experienced
prior manic episodes, a median of 13.5
years had elapsed since the first manic
episode, and a median of 9.5 months had
elapsed since the last manic episode.
Groups were also similar in history of
major depressive episodes (71%), years
since the first depressive episode (me¬
dian, 15.0 years), months since the last
depressive episode (median, 12.5
months), and previous psychiatric hos-
pitalizations (88%).
Among the 146 patients (82%) who
had taken lithium previously, the drug
was reported to have been effective for
the treatment of the most recent prior
episode in 70 patients (48%). However,
it was not tolerated in 13 of these pa¬
tients (Table 1). No significant differ¬
ences were observed among treatment
groups for antimania medications
(lithium, neuroleptic agents, carba-
mazepine, benzodiazepines) adminis¬
tered for 3 or more days within 30 days
before the washout period. From day 7
through day 10, the use of chloral hy¬
drate and lorazepam diminished in both
divalproex- and lithium-treated patients,
whereas use continued at the same level
in placebo-treated patients.
The treatment groups did not differ
at baseline on the primary efficacy vari¬
able or GAS score. The divalproex group
had significantly (P<.05) higher ADRS
depression scores (mean, 7.9±3.3 SD) at
baseline than the lithium group (mean,
6.9±3.1 SD).
By day 8 of treatment, the mean val-
proate and lithium concentrations were
77 µg/mL for valproate and 1 mmol/L
for lithium. Mean dosage peaked on day
18 at slightly more than 2000 mg for
divalproex and 1950 mg for lithium. On
day 21, mean plasma concentrations
were 647 µ /L (93.2 µg/mL) and 1.2
mmol/L, respectively.
Premature Terminations
Sixty-four percent of patients in the
placebo group and 61% in the lithium-
treated group failed to complete all 21
days of treatment compared with only
48% in the divalproex-treated group. The
principal reason for early termination
was lack of efficacy (Table 2). The rate
of termination for lack of efficacy was
significantly lower in the divalproex
group than in the placebo group. The
rate of termination for intolerance to
treatment was higher in the lithium
group than in the placebo group: 11% vs
3%.
SADS-C Results
By day 5, divalproex-treated patients
had greater improvement on the Manic
Syndrome subscale, in terms of statis¬
tical significance, than placebo-treated
patients. By day 10, the same was true
for the primary outcome variable, the
Manic Rating Scale score, and by day 15
for the Behavior and Ideation subscale
score. The differences persisted to the
end of the experimental period (Fig 1).
Figure 2 shows individual change scores
for the three groups. Marked improve¬
ment, defined as at least 50% improve¬
ment in the Manic Syndrome subscale
score, occurred in 48% and 49% of the
divalproex and lithium groups, respec¬
tively, and in only 25% of the placebo
group. The rates of improvement in the
two drug-treated groups were signifi¬
cantly greater than the improvement
observed in the placebo group (P=.004
and P=.025, respectively). Although the
effect sizes for the lithium-treated group
were comparable to those for the dival¬
proex group, most comparisons with pla¬
cebo did not reach significance, a func¬
tion of the smaller size of the lithium
group.
Analysis of individual items on the
Mania Rating scale revealed significantly
greater improvement in divalproex-
treated patients compared with placebo-
treated patients in elevated mood, less
need for sleep, excessive activity, and
motor hyperactivity, and for lithium-
treated patients in excessive activity and
motor hyperactivity. Table 3 shows the
overall changes in Mania Rating scale
scores of the 142 patients for whom base¬
line information was available on respon¬
siveness to lithium for the treatment of
the most recent prior episode. Equiva¬
lent improvement with divalproex oc¬
curred whether or not patients had re¬
sponded to lithium. In contrast, a his¬
tory of previous lithium response was a
strong predictor of response among
lithium-treated patients in this study.
GAS and ADRS Results
Divalproex-treated patients had mar¬
ginally greater improvement on the GAS

Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020

 Table 3.—Changes in Manie Rating Scale Scores of 142 Patients With Known Lithium Responsiveness Prior to the Study*
Treatment Group
Divalproex Lithium Placebo
Previous Lithium -1 - I I
Response No. of Patients Baseline Change No. of Patients Baseline Change No. of Patients Baseline Change
Responders 28 38.2 -7.4 37.6 -15.3 23 39.6 -4.0

Nonresponders 38.6 -1.0 39.1

*0f the 176 patients included in efficacy analyses, 143 had a history of lithium use; one additional patient was excluded since the efficacy of previous lithium treatment could
not be established. Negative values indicate improvement from baseline (see Fig 2).

Table 4.—Treatment-Emergent Adverse Events (COSTART34) Observed in at Least 15% of Any Treatment divalproex-treated patients compared
Group or in Any Group Differing Significantly From Another* with a median of 5% improvement for
Treatment Group, No. (%) the 19 patients treated with placebo.
The current results indicate that dival¬
Divalproex Lithium Placebo
Adverse Event (n=58/69) (n=33/36) (n=58/74) proex is similarly effective in hospital¬
Asthenia 9(13) 7(19)
ized manic patients with and without
7(9)
Constipation 7(10) 6(17) 5(7) rapid-cycling courses and that the re¬
Diarrhea 8(12) 5(14) 13(18) sponse is equivalent in patients previ¬
Dizziness 11 (16) 3(8) 4(5) ously responsive to and nonresponsive
to lithium.
Feverf KD 5(14) 3(4)
Headache 15(22) 14(39) 24 (32)
Divalproex was particularly effective
in improving outcomes on items in the
Nausea 16(23) 11 (31) 11(15) Manic Syndrome subscale. Elevated
Paint 13(19) 1(3) 15(20) mood, less need for sleep, and excessive
Somnolence 13(19) 7(19) 11(15) activity all improved significantly in
Twitching§ 3(8) 0
2(3) divalproex-treated patients. Nurse-
Vomitingil 10(14) 9(25) 3(4) assessed ratings showed similar patterns
*The n's at the top of each column indicate numbers of patients in each treatment group who experienced adverse
of improvement. The ADRS ratings
events over the total in that group. Some patients had more than one adverse event. showed that elation/grandiosity and psy¬
-fLithium > divalproex (Ps.05). chosis improved significantly more in
^Placebo > lithium, divalproex > lithium (Ps.05).
¿Lithium > placebo (Ps.05). divalproex-treated than placebo-treated
IIDivalproex > placebo, lithium > placebo (Ps.05). patients and that mania improved more
in both divalproex-treated and lithium-
than placebo-treated patients; the mean or tended to improve with divalproex treated patients than in those receiving
changes in scores from baseline were treatment. Lithium-treated patients had placebo.
7.6 points in the divalproex group and increased platelet, white blood cell, and Improvements in the Behavior and
3.8 points in the placebo group (P=.06). neutrophil counts. No other clinically Ideation subscale scores and the GAS (a
There were statistically significant dif¬ significant changes in laboratory mea¬ more general measure) were less dra¬
ferences between the divalproex and pla¬ sures were observed in any treatment matic than those in the specific indica¬
cebo groups on three subscales of the group. tors of mania. This may be because ini¬
ADRS, with greater improvement from tial values for these tests were less ab¬
baseline in the divalproex group for COMMENT normal than the baseline scores for the
mania (—4.9 vs -0.2 points, respective¬ This report represents the first large- more specific manic syndrome items;

ly); elation/grandiosity (-2.6 vs -0.7
points, respectively); and psychosis (-2.7
vs +0.6 points, respectively). The lithium
group improved significantly more from
baseline than the placebo group on the
mania subscale only (-5.9 vs -0.2, re¬
spectively).
Safety Data
As displayed in Table 4, significant
treatment group differences were found
only for vomiting (divalproex and lith¬
ium > placebo); fever (lithium > dival¬
proex); general pain, as defined in the
COSTART system34 (divalproex and pla¬
cebo > lithium); and twitching (lithium
> placebo).
Divalproex-treated patients had a
mean reduction in platelet count of
77X10!,/L (32xl09/L to 246X107L), but
no adverse events related to bleeding or
bruising occurred in any patient with a
reduction in platelet count. Indexes of
hepatic function either did not change
scale, controlled evidence that dival¬
proex is an effective treatment for acute
mania. Both divalproex and lithium were
markedly superior to placebo in allevi¬
ating manic symptomatology during the
3-week period of treatment and were
similar in efficacy to one another. The
first test result in the divalproex-treated
patients to show statistically significant
improvement was the Manic Syndrome
subscale score on the fifth day of treat¬
ment, followed by the more global Ma¬
nia Rating scale on the tenth day; these
improvements persisted throughout the
experimental period. Similarly, the per¬
centages of divalproex- and lithium-
treated patients with marked (>50%)
improvement were about twice as great
as in placebo-treated patients. In the
one other placebo-controlled study of
divalproex in mania, which was limited
to patients nonresponsive to, or intol¬
erant of, lithium, Pope et aP reported a
median of 54% improvement in
these measures may have been less sen¬
sitive to short-term changes, or the mea¬
sures may be less specific to manic psy-
chopathology.
Even though improvement was ob¬
served in the majority of patients, a re¬
turn to fully normal functioning within
3 weeks was unusual. However, severely
ill manic patients, when presenting in a
state of psychotic mania, are unlikely to
achieve complete remission within a
3-week treatment period with only one
antimanic drug.14 The patients in this
study did not receive neuroleptic agents
for agitation or psychosis, although these
drugs are usually coadministered with
mood stabilizers in acutely manic pa¬
tients. Also, benzodiazepines and chlor¬
al hydrate were not given after the tenth
day of the study.
Mania, while usually spontaneously
time limited, was not particularly re¬
sponsive to placebo under the conditions
the 17 imposed by the trial. The low placebo

Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020

 response rate is similar to that reported
in two recent placebo-controlled studies
of manic patients.25·35
The time investment and expense in¬
volved in this difficult study limited the
number of patients who could be en¬
rolled. Because lithium was included as
a benchmark treatment to aid in vali¬
dation of study results and not as a pri¬
mary focus of inquiry, the total sample
size for lithium-treated patients was set
at half that of either the divalproex- or
placebo-treated patients. Based on a
treatment-group ratio of 2:1:2 (dival-
proex:lithium:placebo), our best predic¬
tions of power were 92% for divalproex
vs placebo, 95% for lithium vs placebo,
and less than 21% for divalproex vs
lithium. These estimates were based on
the percentage of patients anticipated
to achieve at least 40% improvement
from baseline: 50% in the divalproex
group, 60% in the lithium group, and
20% in the placebo group. Since the pro¬
portions of divalproex- and lithium-
treated patients who improved were so
close, it is doubtful whether a study of
adequate power to examine the differ¬
ences between divalproex and lithium
would be economically feasible.
This is the first placebo-controlled,
parallel-group study of lithium for pa¬
tients with acute mania in the peer-re¬
viewed literature. Previous blinded, pla¬
cebo-controlled studies of lithium in
acute mania were crossover trials19"21
with modest sample sizes (three studies
with a total of 78 patients). Our lithium
response rates, based on the Mania Rat¬
ing scale, were somewhat lower than
those reported in these earlier studies.
The explicit psychometric tools avail¬
able for our trial did not exist at the time
the previously cited studies were per¬
formed. In the 1971 study by Stokes et
al,19 for example, the 75% of patients
reported to have improved included all
patients judged by clinicians to have a
1-point or greater reduction on a 7-point
global mood scale after 10 days of lithium
treatment.
The percentage response we observed
with lithium compares with recent con¬
trolled studies that have reported re¬
sponse rates to lithium of 32%,36 92%,24
and 61%.37 Determination of comparable
response rates in such studies is diffi¬
cult because of varying response crite¬
ria and the reporting of only completers,
which tends to inflate results. For ex¬
ample, Lerer et al37 reported that 11 of
14 subjects responded to lithium, but
their analyses did not include four pa¬
tients whose condition deteriorated or
who had severe adverse effects, inclu¬
sion of which reduces the response rate
to 61%.
Lithium-treated patients had a greater
Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020
proportion of terminations due to adverse
events than patients in the placebo group.
Among side effects affecting at least 15%
of the patient population, only vomiting
was significantly more common in dival¬
proex-treated patients than in patients
treated with placebo. One reason that
divalproex was administered rather than
valproic acid was that it has a lower as¬
sociated incidence of gastrointestinal side
effects.38 Vomiting was more common in
lithium-treated patients than in the other
two groups.
Our data provide detailed informa¬
tion about the baseline characteristics
and symptoms of severe mania in a
large population of patients with bipolar
disorder. The data also demonstrate
which behaviors are sensitive to drug-
induced clinical changes in severely
manic patients and which are not. The
study design provides a useful model
for future prospective pharmacologie tri¬
als in patients with severe manic ill¬
ness.39
The treatment period in our study
was 21 days. We were concerned about
depriving the placebo group of medica¬
tions for a longer time, given that un¬
treated acute mania is so devastating an
illness. Also, based on results reported
by other authors, 21 days seemed to be
the shortest interval that would still per¬
mit unequivocal clinical effects of the
study drugs to become manifest. As a
practical matter in this era of managed
care, hospitalization beyond 3 weeks is
not often authorized. Hence the 3-week
time frame is directly relevant to the
practicing clinician.
Since these patients are extremely ill,
additional effective therapies are needed.
This is especially so since at least 30% to
40% of manic patients do not respond to,
or cannot tolerate, lithium.16·40·41 Also,
controlled and open studies indicate that
lithium is relatively ineffective in cer¬
tain subtypes of bipolar illness: mixed
or dysphoric mania, rapid-cycling vari¬
ants, and mania secondary to organic
illness.6·30·42"44 Previous studies suggest
that valproate or carbamazepine may
be more effective in patients with these
subtypes.23,24·44·45 This is the first placebo-
controlled study to show that divalproex
is equally effective in patients with rapid-
cycling and non-rapid-cycling illness.
This study provides no information
on the usefulness of divalproex as main¬
tenance medication. However, open-
label studies of divalproex administered
as maintenance therapy over longer pe¬
riods have been promising.23
Future investigations are needed to
determine the efficacy and safety of long-
term therapy with divalproex in con¬
trolling the full course of manic-depres¬
sive illness. Another question to be re-
solved concerns the effectiveness of val-
proate for the depressed phase of bipolar
disorder. It also remains to be conclu¬
sively determined whether the use of
concomitant drugs such as benzodiaz-
epines or neuroleptic agents will enhance
the effect of valproate, or whether a
combination of valproate and lithium
would be more effective than either
agent alone.
This study demonstrates that dival¬
proex is effective in relieving the core
symptoms of mania, such as elation, gran¬
diosity, and insomnia. Since the thera¬
peutic response is relatively selective
for core manic symptoms, it is not logi¬
cally attributable to any nonspecific
sedative or activity-blunting effects of
divalproex. The response of divalproex-
treated patients is especially encourag¬
ing considering that only limited use of
adjunctive medications (chloral hydrate
and lorazepam) was allowed in the study,
the patients had chronic and severe ill¬
nesses, and the treatment period was
brief. We therefore recommend dival¬
proex as a safe and effective treatment
for acute manic episodes in patients with
bipolar disorder.
This study was funded in part by a grant from
Abbott Laboratories, North Chicago, 111.
The following institutions and individuals, in ad-
dition to the authors, participated in the Depakote
Mania Study Group program: Abbott Laboratories,
North Chicago, Ill.\p=m-\V.Shu, PhD, P. Johnson,
MBA, M. Blake, MS; University of Texas Health
Science Center at San Antonio.\p=m-\M.Javors, PhD,
L. Ereshefsky, PharmD, T. McLeod; University of
Texas Medical School, Houston.\p=m-\A.Shoaib,
MBBA; Harris County Psychiatric Center, Hous-
ton, Tex.\p=m-\M.Johnson, RN; Department of Psy-
chiatry, Case Western Reserve University School
of Medicine, Cleveland, Ohio.\p=m-\S.Kimmel, MD, A.
Wesley, R. Qualtiere, RN, C. Trivedi; Illinois State
Psychiatric Center, Chicago.\p=m-\J.Javaid, PhD, J.
Peterson; Department ofPsychiatry, University of
Texas Southwestern Medical Center and the Vet-
erans Affairs Medical Center, Dallas.\p=m-\M.Lam-
bert, MD, M. Zielinski, MD, P. Orsulak, PhD, L.
Sharp, RN, L. Akers, MS; Department of Psychia-
try, Indiana University School of Medicine,
Indianapolis.\p=m-\M.Miller, MD, J. Kellams, MD, G.
Woodham, MD; and Larue D. Carter Memorial
Hospital, Indianapolis, Ind.\p=m-\A.Frazer, RN.
Interrater reliability testing for the SADS and
SADS-C rating scales was performed by Edward
G. Altaian, PsyD, at the Illinois State Psychiatric
Institute.
In addition, the patience and dedication shown
by all the nursing personnel involved in this trial
are greatly appreciated by the authors.
References
1. Regier DA, Farmer ME, Rae DS, Locke BZ,
Keith SJ, Judd LL. Comorbidity of mental disor-
ders with alcohol and other drug abuse: results
from the Epidemiologic Catchment Area (ECA)
Study. JAMA. 1990;264:2511-2518.
2. Weissman MM, Leaf PJ, Tischler GL, Blazer
DG, Karno M, Bruce ML. Affective disorders in
five United States communities. Psychol Med. 1988;
18:141-153.
3. Egeland JA. Bipolarity: the iceberg of affective
disorders? Compr Psychiatry. 1983;24:337-344.
4. Altman EG, Janicak PG, Davis JM. Mania: clini-
 cal manifestations and assessment. In: Howells JG,
ed. Modern Perspectives in the Psychiatry of the
Affective Disorders. New York, NY: Brunner/Ma-
zel Inc; 1989:292-302.
5. Gershon ES. Genetics. In: Goodwin FK, Jamison
KR, eds. Manic-Depressive Illness. New York, NY:
Oxford University Press Inc; 1990:369-399.
6. Krauthammer C, Klerman GL. Secondary ma-
nia: manic syndromes associated with antecedent
physical illness or drugs. Arch Gen Psychiatry.
1978;35:1333-1339.
7. Bowden CL, Sarabia F. Diagnosing manic-de-
pressive illness in adolescents. Compr Psychiatry.
1980;21:263-269.
8. Ryan N, Puig-Antich J. Affective illness in ado-
lescence. In: Halls RE, Frances AJ, eds. APA An-
nual Review. Washington, DC: American Psychi-
atric Press; 1986;5:420-450.
9. Weissman MM, Bruce ML, Leaf PJ, Florio LP,
Holzer C. Affective disorders. In: Robins LN, Regier
DA, eds. Psychiatric Disorders in America. New
York, NY: Free Press; 1990:53-80.
10. Mirin SM, Weiss RD, Griffin ML, Michael JL.
Psychopathology in drug abusers and their fami-
lies. Compr Psychiatry. 1991;32:36-51.
11. Ross HE, Glaser FB, Germanson T. The preva-
lence of psychiatric disorders in patients with al-
cohol and other drug problems. Arch Gen Psychia-
try. 1988;45:1023-1031.
12. Coppen A, Standish-Barry H, Bailey J, Hous-
ton G, Silcocks P, Hermon C. Long-term lithium
and mortality. Lancet. 1990;335:1347.
13. SchouM,Juel-Nielsen N, Str\l=o"\mgrenE,Voldby
H. The treatment of manic psychoses by the ad-
ministration of lithium salts. J Neurol Neurosurg
Psychiatry. 1954;17:250-260.
14. Goodwin FK, Jamison KR. Manic-Depressive
Illness. New York, NY: Oxford University Press;
1990:22-25, 36-40, 45-49.
15. Maj M, Pirozzi R, Kemali D. Long-term out-
come prophylaxis in patients initially
of lithium
classified as complete responders. Psychopharma-
cology. 1989;98:535-538.
16. Harrow M, Goldberg JF, Grossman LS, Melt-
zer HY. Outcome in manic disorders: a naturalistic
follow-up study. Arch Gen Psychiatry. 1990;47:665\x=req-\
671.
17. Tohen M, Waternaux CM, Tsuang MT, Hunt
AT. Four-year follow-up of twenty-four first-epi-
sode manic patients. J Affect Disord. 1990;19:79-86.
18. Gelenberg AJ, Kane JM, Keller MB, Lavori P,
Conversions From Système International (SI) Units to Conventional
System* Component
Serum Lithium ion, therapeutic
Blood Thrombocytes (platelets)
*These reference values are not intended to be definitive since each
Downloaded From: https://jamanetwork.com/ by a Infotrieve Inc User on 10/07/2020
Rosenbaum JF, Cole K. Comparison of standard
and low serum levels of lithium for maintenance
treatment of bipolar disorder. N Engl J Med. 1989;
321:1489-1493.
19. Stokes PE, Stoll PM, Shamoian CA, Patton MJ.
Efficacy of lithium as acute treatment of manic\x=req-\
depressive illness. Lancet. 1971;1:1319-1325.
20. Maggs R. Treatment of manic illness with lithium
carbonate. Br J Psychiatry. 1963;109:56-65.
21. Goodwin FK, Murphy DL, Bunney WE.
Lithium-carbonate treatment in depression and ma-
nia. Arch Gen Psychiatry. 1969;21:486-496.
22. Emrich HM, von Zerssen D, Kissling W, M\l=o"\ller
H-J, Windorfer A. Effect of sodium valproate on
mania: the GABA-hypothesis of affective disorder.
Arch Psychiatr Nervenkr. 1980;229:1-16.
23. Calabrese JR, Delucchi GA. Spectrum of effi-
cacy of valproate in 55 patients with rapid-cycling
bipolar disorder. Am J Psychiatry. 1990;147:431\x=req-\
434.
24. Freeman TW, Clothier JL, Pazzaglia P, Lesem
MD, Swann AC. A double-blind comparison of val-
proate and lithium in the treatment of acute mania.
Am J Psychiatry. 1992;149:108-111.
25. Pope HG Jr, McElroy SL, Keck PE Jr, Hudson
JI. Valproate in the treatment of acute mania: a
placebo-controlled study. Arch Gen Psychiatry.
1991;48:62-68.
26. Van Elteren PH. On the combination of inde-
pendent two-sample tests of Wilcoxon. Bull Inst
Intern Statist. 1960;37:351-361.
27. Spitzer RL, Endicott J, Robbins E. Research
Diagnostic Criteria (RDC) for a Selected Group of
Functional Disorders. 2nd ed. New York, NY: New
York State Psychiatric Institute; 1975.
28. Fleiss JL. Statistical Methods for Rates and
Proportions. 2nd ed. New York, NY: John Wiley &
Sons Inc; 1981.
29. Endicott V, Spitzer RL. A diagnostic inter-
view: the schedule for affective disorders and schizo-
phrenia. Arch Gen Psychiatry. 1978;35:837-848.
30. Secunda SK, Katz MM, Swann A, et al. Mania,
diagnosis, state measurement and prediction of
treatment response. J Affect Disord. 1985;8:113\x=req-\
121.
31. Swann AC, Secunda SK, Katz MM, et al. Lithium
treatment of mania: clinical characteristics, speci-
ficity of symptom change, and outcome. Psychiatry
Res. 1986;18:127-141.
32. SAS User's Guide: Statistics, Version 6 Edi-
tion. Cary, NC: SAS Institute Inc; 1991.
Units*(Modified from The SI Manual in Health
SI
Reference SI
Interval* Unit
0.50-1.50 mmol/L
150-450 109/L
laboratory determines its own values. They are provided for illustration only.
33. Mehta CR, Patel NR. A network algorithm for
performing Fisher's exact test in r \m=x\c contingency
tables. J Am Stat Assoc. 1983;78:427-434.
34. COSTART: Coding Symbols for Thesaurus of
Adverse Reaction Terms. 3rd ed. Rockville, Md:
Dept of Health and Human Services, Food and
Drug Administration, Center for Drugs and Bio-
logics, Division of Drug and Biological Product Ex-
perience; 1989.
35. Janicak PG, Sharma RP, Easton M, Cornaty
JE, Davis JM. A double-blind, placebo-controlled
trial of clonidine in the treatment of acute mania.
Psychopharmacol Bull. 1989;25:243-245.
36. Small JG, Klapper MH, Milstein V, et al. Car-
bamazepine compared with lithium in the treat-
ment of mania. Arch Gen Psychiatry. 1991;48:915\x=req-\
921.
37. Lerer B, Moore N, Meyendorff E, Cho SR,
Gerson S. Carbamazepine versus lithium in mania:
a double-blind study. J Clin Psychiatry. 1987;48:
89-93.
38. Wilder BJ, Karas BJ, Penry JK, Asconape J.
Gastrointestinal tolerance of divalproex sodium.
Neurology. 1983;33:808-811.
39. Prien RF, Potter WZ. NIMH workshop report
on treatment of bipolar disorder. Psychopharma-
col Bull. 1990;26:409-427.
40. Prien RF. NIH report: five-center study clari-
fies use of lithium, imipramine for recurrent affec-
tive disorders. Hosp Community Psychiatry. 1984;
35:1097-1098.
41. Small JG, Klapper MH, Kellams JJ, et al. Elec-
troconvulsive therapy compared with lithium in the
management of manic states. Arch Gen Psychiatry.
1988;45:727-732.
42.Kukopulos A, Reginaldi D, Laddumada P, Flo-
risG, Serra G, Tondo L. Course of the manic-de-
pressive cycle and changes caused by treatments.
Neuropsychopharmacology. 1980;13:156-167.
43. Himmelhoch JM, Garfinkel ME. Sources of
lithium resistance in mixed mania. Psychopharma-
col Bull. 1986;22:613-620.
44. Keller MB, Lavori PW, Coryell W, Andreasen
NC, Endicott J, Clayton PJ. Differential outcome of
pure manic, mixed/cycling, and pure depressive epi-
sodes in patients with bipolar illness. JAMA. 1986;
255:3138-3142.
45. Post RM. Introduction: emerging perspectives
on valproate in affective disorders. J Clin Psychia-
try. 1989;50(suppl 3):3-9.
Care)
Conversion Conventional
Factor Reference Conventional
(Divide by) Interval* Unit
1 0.50-1.50 mEq/L
1 150-450 103/cu mm