International Journal of Psychiatry in Clinical Practice, 2005; 9(4): 271 /277
ORIGINAL ARTICLE
Relationship between serum valproate and lithium levels and efficacy and tolerability in bipolar maintenance therapy
PAUL E. KECK JR.1, CHARLES L. BOWDEN3, JANE M. MEINHOLD2, LASZLO GYULAI4, THOMAS J. PRIHODA3, JEFFREY D. BAKER2 & PATRICIA J. WOZNIAK2 1
Psychopharmacology Research Program, Department of Psychiatry, University of Cincinnati College of Medicine and Mental Health Care Line and General Clinical Research Center, Veterans Affairs Medical Center, Cincinnati, OH, 2Abbott Laboratories, Abbott Park, IL, 3Department of Psychiatry, University of Texas Health Science Center, San Antonio, TX, and 4 Department of Psychiatry, University of Pennsylvania Medical Center, Philadelphia, PA, USA
Abstract Background. In spite of widespread recognition of the importance of maintenance treatment for bipolar disorder, there are relatively few available well-designed clinical trials that have provided rigorous evidence for the efficacy of specific agents. One approach used effectively in analyses of lithium studies has been the stratification of efficacy and tolerability results based on serum drug concentrations. Therefore, we conducted a similar analysis of the efficacy of divalproex or lithium, based on serum concentrations in the maintenance treatment of bipolar I disorder in a recent 12-month placebo-controlled trial. Methods. This was a post-hoc analysis of results obtained in a 12-month, double-blind, placebo-controlled trial of divalproex and lithium involving 372 patients (intent-to-treat). The patient set was stratified into four therapeutic drug concentration ranges (Non-therapeutic, Low Therapeutic, Medium Therapeutic, and High Therapeutic) for both divalproex and lithium. Efficacy measures were Kaplan /Meier survival analyses of time to discontinuation for any reason, time to discontinuation for a protocol-defined manic episode, and time to discontinuation for a manic or depressive episode. Results. Significant differences between divalproex at the Medium Therapeutic range (75 /99.9 mg/ml) and placebo were demonstrated in Kaplan /Meier survival results for discontinuation for any reason (median survival time: divalproex, 8 months; placebo, 4 months; P B/0.05) and for discontinuation for a manic or depressive episode (median survival time: divalproex, 8 months; placebo, 3 months, P /0.003). At 12 months, the proportion of divalproex-treated patients (Medium Therapeutic range) who did not discontinue for a protocol-defined manic episode (85%) was higher than the proportion of lithium (Medium Therapeutic range; 70%) or placebo-treated (83%) patients. Conclusions. Divalproex at the Medium Therapeutic range provided significantly better bipolar maintenance treatment response than placebo in survival analyses, suggesting a possible serum concentration target range for clinicians in providing optimal treatment response. The value of this analytic approach, used for the first time here for divalproex, is discussed, along with a call for further research into optimal therapeutic drug levels.
Key Words: Valproate, lithium, bipolar maintenance
Introduction In studies of pharmacological treatment of bipolar disorder, treatment of acute manic or mixed episodes has generally been used to assess the efficacy and tolerability of specific agents. However, recent studies of the course of bipolar disorder over time have confirmed the clinical impression that patients spend far more time in intercurrent periods than in acute episodes that meet criteria for mania or major depression. Recent studies have found that in both bipolar I and II disorder, patients were asymptomatic approximately half the time. Roughly 75% of symptomatic periods were characterized by subsyn-
dromal dysthymia, minor depression, or hypomania, rather than full-blown mania or major depression [1,2]. The American Psychiatric Association Practice Guidelines revision recommends initiation of maintenance treatment following an initial manic episode: to prevent relapse; to reduce the risk of suicide, the frequency of cycling, and the severity of subthreshold symptoms; and to facilitate functional improvement [3]. However, there are few well-designed, randomized, controlled trials examining the efficacy of putative mood-stabilizing agents in the maintenance therapy of bipolar disorder.
Correspondence: Paul E. Keck Jr., MD, Department of Psychiatry, University of Cincinnati College of Medicine, 231 Albert Sabin Way, PO Box 670559, Cincinnati, OH 45267-0559, USA. Tel: /1 513 558 8626. Fax: /1 513 558 6131. E-mail: [email protected]
(Received 17 February 2005; accepted 7 July 2005) ISSN 1365-1501 print/ISSN 1471-1788 online # 2005 Taylor & Francis DOI: 10.1080/13651500500305622
272
P. E. Keck et al.
Early placebo-controlled studies of bipolar maintenance using lithium were limited by a number of design flaws, including poorly-defined diagnostic criteria for study inclusion, inadequate characterization and description of reasons for premature discontinuation, and analyses based only on patients who completed the trials [3]. In addition, the frequent use of study designs in which patients assigned to placebo were abruptly discontinued from lithium may have led to artificially elevated rates of placebo relapse, creating a bias toward treatment response to lithium [4]. More recent open-label and placebo-controlled studies of lithium have produced varying estimates of its efficacy in maintenance therapy. For example, in one large 18-month placebo-controlled trials in which lithium was used as a comparator to lamotrigine, lithium was associated with a highly significant (P /0.003) extension of time to intervention for a mood episode compared with placebo [5]. In contrast, a 12-month study of divalproex or lithium, which required a full syndromal mood episode to establish maintenance treatment failure reported no significant difference between lithium and placebo [6]. To some extent, the difficulties in interpreting the results of trials of lithium or other mood stabilizers stem from inherent complexities in the design of bipolar maintenance studies. The selection of clinically meaningful, standardized outcome indices that reflect the time element, the risk of both depressive and manic relapse, as well as the development of subthreshold symptoms, is challenging. In addition, the high degree of risk (including that of suicide) associated with full syndromal relapse introduces ethical dilemmas in the conduct of placebo-controlled studies, often leading to the selection of patients with milder, rather than more severe, forms of illness [3,7]. One method of analyzing the effectiveness of maintenance therapy is by stratifying patient response based on serum drug concentrations. This approach has been somewhat successful in identifying a therapeutic ‘‘window’’ of serum lithium concentrations, with an upper bound defined by tolerability and the risk of adverse events and a lower bound defined by unacceptable reduction in efficacy [8 /10]. The efficacy of divalproex in attenuating or eliminating acute manic symptoms has been clearly established in double-blind, placebo-controlled trials, and is comparable to that of lithium [11,12]; open-label and open randomized studies have also suggested that valproate can reduce the severity and frequency of recurrent manic and depressive episodes [13/15]. Despite these positive reports, a recent 12-month placebo-controlled, double-blind study of the efficacy of divalproex in maintenance treatment of bipolar disorder failed to demonstrate
statistically significant improvement on primary efficacy measures including time to the development of any mood episode, time to the occurrence of a protocol-defined manic episode, or time to the development of depression [6]. However, divalproex treatment was significantly superior to placebo on most secondary efficacy measures and to lithium on several secondary measures. We therefore hypothesized that stratification by serum drug concentration might reveal treatment effects not discernable in the full treatment groups. We report here the results of the 12-month study in which we assessed treatment efficacy and tolerability based on serum level ranges of concentrations of divalproex and lithium. Methods The present report is a post-hoc analysis of data obtained during a 12-month randomized, placebocontrolled, parallel-group study of the efficacy of divalproex sodium and lithium in maintenance therapy for bipolar disorder, conducted at 37 study centers in the United States. The methodology used in this study has been previously described in detail [6]. Patient selection Subjects included adults aged 18 /75 years meeting the DSM-III-R criteria for bipolar disorder [16], who had experienced an index manic episode as diagnosed using the Structured Clinical Interview for DSM-III-R [17], and who had experienced at least one other manic episode within the preceding 3 years. Enrollment in an initial open-label phase was permitted while the patient was manic, partially recovered from a manic episode, or euthymic after the episode, but not while the patient was depressed. Patients were excluded if they had a history of intolerance to divalproex or lithium; alcohol abuse within the previous 6 months; current substance dependence or positive results on a urine toxicology test; concomitant drug therapy with the potential to confound study results; central nervous system, neuromuscular, or uncontrolled systemic disorders; serious suicide risk; ongoing individual psychotherapy; or evidence of failure to adhere to the openphase protocol. Pregnant female patients were also excluded. Study design The 52-week double-blind study period was preceded by an initial open-label phase lasting 3 months or less, during which the index manic episode was treated at the discretion of the investigator (depot neuroleptics or electroconvulsive therapy were not permitted). Patients were excluded from entry into the double-blind study, and referred for treatment, if
Serum valproate and lithium levels in bipolar maintenance therapy they required hospitalization for depression or mania, had a Mania Rating Scale (MRS) score of /16, found side effects intolerable, had evidence of medical danger, became pregnant, or had a score ]/4 on the suicide item from the Schedule for Affective Disorders and Schizophrenia (SADS). Patients who scored 5/11 on the MRS, 5/13 on the Depressive Syndrome Scale (DSS), and /60 on the Global Assessment Scale (GAS) on two consecutive occasions at least 6 days apart were randomized into the double-blind phase for treatment with divalproex sodium, lithium carbonate, or placebo. Randomization occurred on the day of the second evaluation meeting the above criteria. All patients provided written consent and the protocol was approved by the institutional review boards of the respective institutions. Patients were randomized to receive divalproex, lithium, or placebo in a 2:1:1 ratio. Any patients who were taking divalproex or lithium on the day of randomization were tapered gradually during the initial 2 weeks of the maintenance treatment phase; if a patient was taking both drugs, one was discontinued before randomization. All study drugs were administered 3 times daily. The daily dose was adjusted upward from the starting dose, based on body weight, to maintain target serum trough drug concentrations of 71 /125 mg/ml for valproate or 0.8 /1.2 mmol/l for lithium as tolerated. During the initial 12 weeks of maintenance treatment, scheduled study visits occurred weekly or biweekly, and monthly thereafter. Serum concentrations of study drug were assessed at weeks 2, 3, 4, 8, 12, 16, 20, 24, 28, 36, 44, and week 52 or final visit. To minimize the recurrence of manic symptoms during withdrawal of open-label medication, lorazepam, and if necessary, haloperidol, were initially permitted as rescue medications. One or two courses of lorazepam were permitted, including a 14-day maximum course (dose 5/6 mg/day) during the initial month of maintenance treatment and a 7-day maximum course during the duration of the study. Haloperidol (dose 5/10 mg/day) was permitted during the second consecutive week of lorazepam use in the first month only. Neither drug could be administered within 8 hours prior to behavioral assessments.
273
99.9 mg/ml; and high therapeutic, /100.0 mg/ml. For lithium, the corresponding serum level ranges were: non-therapeutic, B/0.49 mmol/l; low therapeutic, 0.50 /0.79 mmol/l; medium therapeutic, 0.80 /1.00 mmol/l; high therapeutic, /1.00 mmol/l. Within each defined therapeutic category, Kaplan /Meier Survival and Cox regression analyses [18] were used to determine time to study discontinuation. Separate analyses were performed based on discontinuation for any reason, discontinuation for a protocol-defined manic episode (manic episode accompanied by an MRS score /16, or one requiring hospitalization), or discontinuation for mania or depression (protocol-defined manic or depressive episode requiring the use of an antidepressant or premature discontinuation because of depressive symptoms). All tests were two-tailed; tests of significance have been previously described [6].
Results Demographic and clinical characteristics at baseline A total of 571 patients were enrolled in the openphase study period; of these, 372 (divalproex, N /187; lithium, N /91; placebo, N /94) were subsequently randomized into the double-blind maintenance treatment phase. The treatment groups were well matched, with no significant differences with respect to age, proportion of male and female patients, or race. Mean MRS, DSS, and GAS scores; mean age of first manic episode; and mean age of first depressive episode were all similar between groups. Approximately half of the patients in each treatment group reported /10 prior manic episodes and/or /10 depressive episodes. Serum drug level measurements sufficient for analysis were available for a total of 331 patients (divalproex, N /169; lithium, N /74; placebo, N /88), who constituted the study group for the present study. The stratification by therapeutic category is shown in Table I. With regard to study discontinuation for any reason, Kaplan/Meier survival curves for divalproex and lithium are shown in Figure 1. The time in maintenance treatment before discontinuation for Table I. Patient stratification by therapeutic category.
Analysis Serum drug concentration
Based on a retrospective analysis of individual patients’ serum drug level ranges, randomized patients who received study drug were stratified into four categories (non-therapeutic, low therapeutic, medium therapeutic, and high therapeutic) for each drug treatment. For divalproex, the valproate serum level ranges used to define these categories were: non-therapeutic, B/49.9 mg/ml; low therapeutic, 50.0 /74.9 mg/ml; medium therapeutic, 75.0 /
Therapeutic category
Divalproex, range (mg/ml)
Lithium, range (mmol/l)
Non-therapeutic B/49.9; N /11 B/0.49; N /5 Low therapeutic 50.0 /74.9; N /45 0.50 /0.79; N /16 0.80 /1.00; N /37 Medium 75.0 /99.9; N /85 therapeutic High Therapeutic /100.0; N /28 /1.00; N /16 N / number of patients. Placebo: N /88.
274
P. E. Keck et al. A
100% 80%
Survival
Placebo
60%
Nontherapeutic
*
Low therapeutic
40%
Medium therapeutic High therapeutic
20% 0% 0
1
2
3
4
5
6
7
8
9
10
11 12
Months of Study
B
100% 80%
Survival
Placebo Nontherapeutic
60%
Low therapeutic
40%
Medium therapeutic High therapeutic
20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study Figure 1. Survival analysis: discontinuation for any reason. Kaplan /Meier survival curves for divalproex-treated patients (A) and lithiumtreated patients (B) who were discontinued from the double-blind treatment period of the study for any reason. Only divalproex at the Medium Therapeutic concentration range (*) demonstrated a statistically significant difference (P B/0.05) from placebo. *Significant difference from placebo (P B/0.05).
any reason (i.e. a mood episode, or early drop out for any reason: medication intolerance, non-compliance to protocol, moved, lost to follow-up) was significantly greater for the divalproex treatment group in the Medium Therapeutic serum concentration range (valproate concentration 75 /99.9 mg/ml) than for placebo (P B/0.05). Median survival time for the Medium Therapeutic divalproex group was 8 months, compared with 4 months for placebo. Among lithium-treated patients, no lithium serum level treatment group differed significantly from placebo. The number of patients who relapsed to protocoldefined mania was too small to permit Kaplan/ Meier significance analysis. However, in comparing all treatments, divalproex treatment in the Medium Therapeutic category demonstrated the highest proportion of patients (85%) who survived to month 8. There were no indications of numerical superiority for any of the lithium serum level treatment groups. Kaplan /Meier survival curves for divalproex and lithium with regard to study discontinuation for mania or depression are shown in Figure 2. The only statistically significant difference from placebo observed for divalproex was for the Medium Therapeutic serum concentration range group. Median survival time for this group was 8 months, compared with 3 months for placebo. No lithium serum level group differed from placebo on this analysis. Median
survival times for placebo, lithium, and divalproex groups are summarized in Table II. Discussion This is the first published study that examines response to divalproex maintenance treatment of bipolar disorder as a function of serum drug concentrations. Significant improvement was observed for patients in the divalproex Medium Therapeutic category (serum valproate concentrations 75.0 /99.9 mg/ml) with respect to two outcome measures: discontinuation for any reason, and discontinuation for mania or depression. The number of patients who experienced the third outcome measure, time to protocol-defined mania, was too small to permit adequate statistical analysis, but appears numerically to support the efficacy of this valproate serum concentration range for delaying time to mania. The significant improvement with regard to discontinuation for any reason can be taken as an overall measure of treatment effectiveness, incorporating both elements of efficacy and drug tolerability. This result suggests that for divalproex the Medium Therapeutic range provides optimal effectiveness in preventing relapse while achieving acceptable tolerability during maintenance therapy in a subgroup of recently manic bipolar 1 patients who may not have a severe form of the disorder [6]. Whether such a
Serum valproate and lithium levels in bipolar maintenance therapy
275
A 100%
* 80% Placebo Nontherapeutic Low therapeutic Medium therapeutic High therapeutic
Survival
60% 40% 20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study B
100% 80% Placebo Nontherapeutic Low therapeutic Medium therapeutic High therapeutic
Survival
60% 40% 20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study
Figure 2. Survival analysis: discontinuation for mania or depression. Kaplan /Meier survival curves for divalproex-treated patients (A) and lithium-treated patients (B) who were discontinued from the double-blind treatment period of the study for mania or depression (protocoldefined mania or depressive episode requiring the use of an antidepressant or premature discontinuation because of depressive symptoms). Only divalproex at the Medium Therapeutic concentration range demonstrated a statistically significant difference (P /0.003) from placebo. *Significant difference from placebo (P /0.003).
serum range would also apply to other subgroups of bipolar patients in maintenance treatment remains to be determined. A comparison of Kaplan /Meier survival curves for discontinuation resulting from protocol-defined mania (not shown) and for discontinuation resulting from mania or depression strongly suggests that relapse into depressive mood episodes is a more significant concern during bipolar maintenance therapy than relapse into mania. An analysis of the original study data with regard to prevention of bipolar depression has recently been published [19]. This study showed that divalproex was more effective than lithium in delaying time to depressive relapse among patients who had responded to Table II. Median survival time: Discontinuation for protocoldefined mania or depression. Treatment group (N )
Placebo (88) Lithium Nontherapeutic (5) Low therapeutic (16) Medium Therapeutic (37) High Therapeutic (16) Divalproex Nontherapeutic (11) Low Therapeutic (45) Medium Therapeutic (85) High Therapeutic (28) NC, not calculable.
Median survival time, months (95% confidence limit) 3.5 (2, 5) 1 4 4 3
(1, (2, (2, (1,
3) 10) 9) 5)
5 5 8 2
(1, (4, (5, (1,
NC) 10) 12) 3)
divalproex during the open-label period, and among those who had previously been hospitalized (suggesting greater illness severity). Divalproex treatment was also associated with less deterioration than lithium with respect to depressive symptoms. In addition, the combination of an SSRI with divalproex was significantly more effective in treating breakthrough depression than an SSRI alone (SSRI/placebo) [19]. Between-group differences with regard to discontinuation for depression alone did not reach statistical significance in this study. However, the significant improvement associated with divalproex in the Medium Therapeutic range with regard to discontinuation for mania or depression, together with the predominance of depression relative to mania as a cause for discontinuation, suggest some efficacy for divalproex at Medium Therapeutic serum drug concentrations in preventing depressive relapse. This study had the advantage of relatively frequent serum level monitoring over a long period of time. Additionally, efforts to obtain serum levels during essentially trough serum level time periods, and analysis of serum levels by a central lab may have strengthened the results. We averaged serum levels over the time of activity in the study. It is possible that this strategy failed to detect outcome differences related to visit-by-visit fluctuations in serum levels. However, our analyses of stability of serum levels within patients suggest that central tendency of
276
P. E. Keck et al.
serum levels was maintained across time for both divalproex and lithium treated patients. Certain limitations in the original study have been addressed in previous publications [6,7,19,20]. The failure to demonstrate significant improvement in the primary outcome measures may have reflected the stringency of the a priori criteria for depressive or manic relapse. This definition in part contributed to a lower than predicted number of relapse events, reducing the power of the statistical analyses [6]. An additional contributing factor may have been the low rate of relapse in the placebo group relative to earlier placebo-controlled double-blind studies of lithium, potentially reducing the effect size. The low rate of placebo relapse may also have been a function of the study duration, as placebo-treated patients in an earlier lithium trial continued to relapse after 12 months [21]. The failure to demonstrate treatment efficacy for any dose of lithium in this study is consistent with the original study, and may also reflect to some extent the low rate of placebo relapse [6]. In addition, the number of subjects in the lithium arm of the study, which was included principally as an active control group, was essentially half of the number in the divalproex group, reducing power. Because the efficacy of lithium in bipolar maintenance treatment continues to be supported by recent trials that were not compromised by earlier methodological failings [5,8,22,23], these results should be interpreted with great caution. However, the results do not tend to indicate that one specific serum lithium range, such as that which is established as effective for treatment of acute mania, is overall more effective than other serum level ranges. Because the linkage of lithium responsiveness in maintenance to serum lithium levels remains a matter of controversy, these results remain of interest in that they would seem to support the case of those authorities who recommend maintenance treatment with lithium at lower than 0.8 mEq/l for some patients. This study was also limited by the relatively small number of patients within the stratified therapeutic categories, reducing the statistical power of the applied analyses. The use of discontinuation for any reason as a measure of tolerability is to some extent compromised by the inclusion of relapserelated discontinuations in this variable. It would be desirable in future studies to provide a separate analysis of discontinuations related to adverse events. Notwithstanding these limitations, this study demonstrates a significant treatment effect for divalproex, when dosed to achieve serum valproate concentrations from 75.0 to 99.9 mg/ml, in the prevention of relapse during maintenance therapy of bipolar disorder I. The novel use of serum valproate concentrations to assess efficacy may help clinicians identify treatment targets for bipolar
maintenance, and establishes a new direction for further research into the effectiveness of divalproex in this challenging therapeutic arena. Key points . Retrospective analysis of 12-month, doubleblind, placebo-controlled trial of divalproex sodium and lithium involving 372 patients (intent-to-treat) . Efficacy measures were Kaplan /Meier survival analyses of time to discontinuation for any reason, time to discontinuation for a protocoldefined manic episode, and time to discontinuation for a manic or depressive episode . Serum valproate levels of 75/99.9 mg/ml were most effective in preventing relapse to mania or depression and discontinuation for any reason. The number of patients who relapsed to protocol-defined mania was too small to permit adequate statistical analysis Statement of interest The authors have no conflict of interest with any commercial or other associations in connection with the submitted article.
References [1] Judd LL, Akiskal HS, Schettler PJ, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry 2003;60:261 /9. [2] Judd LL, Akiskal HS, Schettler PJ, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry 2002;59:530 /7. [3] Practice guideline for the treatment of patients with bipolar disorder (revision). Am J Psychiatry 2002;159:1 /50. [4] Suppes T, Baldessarini RJ, Faedda GL, et al. Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Arch Gen Psychiatry 1991;48:1082 /8. [5] Bowden CL, Calabrese JR, Sachs G, et al. A placebocontrolled 18-month trial of lamotrigine and lithium maintenance treatment in recently manic or hypomanic patients with bipolar I disorder. Arch Gen Psychiatry 2003;60:392 / 400. [6] Bowden CL, Calabrese JR, McElroy SL, et al. A randomized, placebo-controlled 12-month trial of divalproex and lithium in treatment of outpatients with bipolar I disorder. Divalproex Maintenance Study Group. Arch Gen Psychiatry 2000;57:481 /9. [7] Bowden CL, Swann AC, Calabrese JR, et al. Maintenance clinical trials in bipolar disorder: Design implications of the divalproex-lithium-placebo study. Psychopharmacol Bull 1997;33:693 /9. [8] Gelenberg AJ, Kane JM, Keller MB, et al. Comparison of standard and low serum levels of lithium for maintenance treatment of bipolar disorder. New Engl J Med 1989;321: 1489 /93. [9] Hopkins HS, Gelenberg AJ. Serum lithium levels and the outcome of maintenance therapy of bipolar disorder. Bipolar Disorder 2000;2:174 /9. /
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
277
Serum valproate and lithium levels in bipolar maintenance therapy [10] Sproule B. Lithium in bipolar disorder: can drug concentrations predict therapeutic effect? Clin Pharmacokinet 2002; 41:639 /60. [11] Bowden CL, Brugger AM, Swann AC, et al. Efficacy of divalproex vs lithium and placebo in the treatment of mania. The Depakote Mania Study Group. J Am Med Assoc 1994; 271:918 /24. [12] Pope HG Jr, McElroy SL, Keck PE Jr, et al. Valproate in the treatment of acute mania. A placebo-controlled study. Arch Gen Psychiatry 1991;48:62 /8. [13] Calabrese JR, Delucchi GA. Spectrum of efficacy of valproate in 55 patients with rapid-cycling bipolar disorder. Am J Psychiatry 1990;147:431 /4. [14] McElroy SL, Keck PE Jr, Pope HG Jr. Sodium valproate: its use in primary psychiatric disorders. J Clin Psychopharmacol 1987;7:16 /24. [15] Lambert P, Venaud G. Comparative study of valpromide versus lithium in the treatment of affective disorders. Nervure 1992;5:57 /65. [16] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. Revised 3rd ed. Washington, DC: American Psychiatric Association; 1987. [17] Spitzer R, Williams J, Gibbons M, et al. Structured clinical interview for the DSM-III-R. Washington, DC: American Psychiatric Press Inc; 1990. /
/
/
/
/
/
/
/
/
/
/
/
[18] Lawless J. Statistical models and methods for lifetime data. New York: John Wiley & Sons; 1982. [19] Gyulai L, Bowden CL, McElroy SL, et al. Maintenance efficacy of divalproex in the prevention of bipolar depression. Neuropsychopharmacology 2003;28:1374 /82. [20] Baldessarini RJ, Tohen M, Tondo L. Maintenance treatment in bipolar disorder. Arch Gen Psychiatry 2000;57: 490 /2. [21] Prien RF, Caffey EM Jr, Klett CJ. Factors associated with treatment success in lithium carbonate prophylaxis. Report of the Veterans Administration and National Institute of Mental Health collaborative study group. Arch Gen Psychiatry 1974;31:189 /92. [22] Tondo L, Baldessarini RJ, Hennen J, et al. Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders. Am J Psychiatry 1998;155:638 / 45. [23] Berghofer A, Kossmann B, Muller-Oerlinghausen B. Course of illness and pattern of recurrences in patients with affective disorders during long-term lithium prophylaxis: A retrospective analysis over 15 years. Acta Psychiatr Scand 1996; 93:349 /54. /
/
/
/
/
/
/
/
/
/
ORIGINAL ARTICLE
Relationship between serum valproate and lithium levels and efficacy and tolerability in bipolar maintenance therapy
PAUL E. KECK JR.1, CHARLES L. BOWDEN3, JANE M. MEINHOLD2, LASZLO GYULAI4, THOMAS J. PRIHODA3, JEFFREY D. BAKER2 & PATRICIA J. WOZNIAK2 1
Psychopharmacology Research Program, Department of Psychiatry, University of Cincinnati College of Medicine and Mental Health Care Line and General Clinical Research Center, Veterans Affairs Medical Center, Cincinnati, OH, 2Abbott Laboratories, Abbott Park, IL, 3Department of Psychiatry, University of Texas Health Science Center, San Antonio, TX, and 4 Department of Psychiatry, University of Pennsylvania Medical Center, Philadelphia, PA, USA
Abstract Background. In spite of widespread recognition of the importance of maintenance treatment for bipolar disorder, there are relatively few available well-designed clinical trials that have provided rigorous evidence for the efficacy of specific agents. One approach used effectively in analyses of lithium studies has been the stratification of efficacy and tolerability results based on serum drug concentrations. Therefore, we conducted a similar analysis of the efficacy of divalproex or lithium, based on serum concentrations in the maintenance treatment of bipolar I disorder in a recent 12-month placebo-controlled trial. Methods. This was a post-hoc analysis of results obtained in a 12-month, double-blind, placebo-controlled trial of divalproex and lithium involving 372 patients (intent-to-treat). The patient set was stratified into four therapeutic drug concentration ranges (Non-therapeutic, Low Therapeutic, Medium Therapeutic, and High Therapeutic) for both divalproex and lithium. Efficacy measures were Kaplan /Meier survival analyses of time to discontinuation for any reason, time to discontinuation for a protocol-defined manic episode, and time to discontinuation for a manic or depressive episode. Results. Significant differences between divalproex at the Medium Therapeutic range (75 /99.9 mg/ml) and placebo were demonstrated in Kaplan /Meier survival results for discontinuation for any reason (median survival time: divalproex, 8 months; placebo, 4 months; P B/0.05) and for discontinuation for a manic or depressive episode (median survival time: divalproex, 8 months; placebo, 3 months, P /0.003). At 12 months, the proportion of divalproex-treated patients (Medium Therapeutic range) who did not discontinue for a protocol-defined manic episode (85%) was higher than the proportion of lithium (Medium Therapeutic range; 70%) or placebo-treated (83%) patients. Conclusions. Divalproex at the Medium Therapeutic range provided significantly better bipolar maintenance treatment response than placebo in survival analyses, suggesting a possible serum concentration target range for clinicians in providing optimal treatment response. The value of this analytic approach, used for the first time here for divalproex, is discussed, along with a call for further research into optimal therapeutic drug levels.
Key Words: Valproate, lithium, bipolar maintenance
Introduction In studies of pharmacological treatment of bipolar disorder, treatment of acute manic or mixed episodes has generally been used to assess the efficacy and tolerability of specific agents. However, recent studies of the course of bipolar disorder over time have confirmed the clinical impression that patients spend far more time in intercurrent periods than in acute episodes that meet criteria for mania or major depression. Recent studies have found that in both bipolar I and II disorder, patients were asymptomatic approximately half the time. Roughly 75% of symptomatic periods were characterized by subsyn-
dromal dysthymia, minor depression, or hypomania, rather than full-blown mania or major depression [1,2]. The American Psychiatric Association Practice Guidelines revision recommends initiation of maintenance treatment following an initial manic episode: to prevent relapse; to reduce the risk of suicide, the frequency of cycling, and the severity of subthreshold symptoms; and to facilitate functional improvement [3]. However, there are few well-designed, randomized, controlled trials examining the efficacy of putative mood-stabilizing agents in the maintenance therapy of bipolar disorder.
Correspondence: Paul E. Keck Jr., MD, Department of Psychiatry, University of Cincinnati College of Medicine, 231 Albert Sabin Way, PO Box 670559, Cincinnati, OH 45267-0559, USA. Tel: /1 513 558 8626. Fax: /1 513 558 6131. E-mail: [email protected]
(Received 17 February 2005; accepted 7 July 2005) ISSN 1365-1501 print/ISSN 1471-1788 online # 2005 Taylor & Francis DOI: 10.1080/13651500500305622
272
P. E. Keck et al.
Early placebo-controlled studies of bipolar maintenance using lithium were limited by a number of design flaws, including poorly-defined diagnostic criteria for study inclusion, inadequate characterization and description of reasons for premature discontinuation, and analyses based only on patients who completed the trials [3]. In addition, the frequent use of study designs in which patients assigned to placebo were abruptly discontinued from lithium may have led to artificially elevated rates of placebo relapse, creating a bias toward treatment response to lithium [4]. More recent open-label and placebo-controlled studies of lithium have produced varying estimates of its efficacy in maintenance therapy. For example, in one large 18-month placebo-controlled trials in which lithium was used as a comparator to lamotrigine, lithium was associated with a highly significant (P /0.003) extension of time to intervention for a mood episode compared with placebo [5]. In contrast, a 12-month study of divalproex or lithium, which required a full syndromal mood episode to establish maintenance treatment failure reported no significant difference between lithium and placebo [6]. To some extent, the difficulties in interpreting the results of trials of lithium or other mood stabilizers stem from inherent complexities in the design of bipolar maintenance studies. The selection of clinically meaningful, standardized outcome indices that reflect the time element, the risk of both depressive and manic relapse, as well as the development of subthreshold symptoms, is challenging. In addition, the high degree of risk (including that of suicide) associated with full syndromal relapse introduces ethical dilemmas in the conduct of placebo-controlled studies, often leading to the selection of patients with milder, rather than more severe, forms of illness [3,7]. One method of analyzing the effectiveness of maintenance therapy is by stratifying patient response based on serum drug concentrations. This approach has been somewhat successful in identifying a therapeutic ‘‘window’’ of serum lithium concentrations, with an upper bound defined by tolerability and the risk of adverse events and a lower bound defined by unacceptable reduction in efficacy [8 /10]. The efficacy of divalproex in attenuating or eliminating acute manic symptoms has been clearly established in double-blind, placebo-controlled trials, and is comparable to that of lithium [11,12]; open-label and open randomized studies have also suggested that valproate can reduce the severity and frequency of recurrent manic and depressive episodes [13/15]. Despite these positive reports, a recent 12-month placebo-controlled, double-blind study of the efficacy of divalproex in maintenance treatment of bipolar disorder failed to demonstrate
statistically significant improvement on primary efficacy measures including time to the development of any mood episode, time to the occurrence of a protocol-defined manic episode, or time to the development of depression [6]. However, divalproex treatment was significantly superior to placebo on most secondary efficacy measures and to lithium on several secondary measures. We therefore hypothesized that stratification by serum drug concentration might reveal treatment effects not discernable in the full treatment groups. We report here the results of the 12-month study in which we assessed treatment efficacy and tolerability based on serum level ranges of concentrations of divalproex and lithium. Methods The present report is a post-hoc analysis of data obtained during a 12-month randomized, placebocontrolled, parallel-group study of the efficacy of divalproex sodium and lithium in maintenance therapy for bipolar disorder, conducted at 37 study centers in the United States. The methodology used in this study has been previously described in detail [6]. Patient selection Subjects included adults aged 18 /75 years meeting the DSM-III-R criteria for bipolar disorder [16], who had experienced an index manic episode as diagnosed using the Structured Clinical Interview for DSM-III-R [17], and who had experienced at least one other manic episode within the preceding 3 years. Enrollment in an initial open-label phase was permitted while the patient was manic, partially recovered from a manic episode, or euthymic after the episode, but not while the patient was depressed. Patients were excluded if they had a history of intolerance to divalproex or lithium; alcohol abuse within the previous 6 months; current substance dependence or positive results on a urine toxicology test; concomitant drug therapy with the potential to confound study results; central nervous system, neuromuscular, or uncontrolled systemic disorders; serious suicide risk; ongoing individual psychotherapy; or evidence of failure to adhere to the openphase protocol. Pregnant female patients were also excluded. Study design The 52-week double-blind study period was preceded by an initial open-label phase lasting 3 months or less, during which the index manic episode was treated at the discretion of the investigator (depot neuroleptics or electroconvulsive therapy were not permitted). Patients were excluded from entry into the double-blind study, and referred for treatment, if
Serum valproate and lithium levels in bipolar maintenance therapy they required hospitalization for depression or mania, had a Mania Rating Scale (MRS) score of /16, found side effects intolerable, had evidence of medical danger, became pregnant, or had a score ]/4 on the suicide item from the Schedule for Affective Disorders and Schizophrenia (SADS). Patients who scored 5/11 on the MRS, 5/13 on the Depressive Syndrome Scale (DSS), and /60 on the Global Assessment Scale (GAS) on two consecutive occasions at least 6 days apart were randomized into the double-blind phase for treatment with divalproex sodium, lithium carbonate, or placebo. Randomization occurred on the day of the second evaluation meeting the above criteria. All patients provided written consent and the protocol was approved by the institutional review boards of the respective institutions. Patients were randomized to receive divalproex, lithium, or placebo in a 2:1:1 ratio. Any patients who were taking divalproex or lithium on the day of randomization were tapered gradually during the initial 2 weeks of the maintenance treatment phase; if a patient was taking both drugs, one was discontinued before randomization. All study drugs were administered 3 times daily. The daily dose was adjusted upward from the starting dose, based on body weight, to maintain target serum trough drug concentrations of 71 /125 mg/ml for valproate or 0.8 /1.2 mmol/l for lithium as tolerated. During the initial 12 weeks of maintenance treatment, scheduled study visits occurred weekly or biweekly, and monthly thereafter. Serum concentrations of study drug were assessed at weeks 2, 3, 4, 8, 12, 16, 20, 24, 28, 36, 44, and week 52 or final visit. To minimize the recurrence of manic symptoms during withdrawal of open-label medication, lorazepam, and if necessary, haloperidol, were initially permitted as rescue medications. One or two courses of lorazepam were permitted, including a 14-day maximum course (dose 5/6 mg/day) during the initial month of maintenance treatment and a 7-day maximum course during the duration of the study. Haloperidol (dose 5/10 mg/day) was permitted during the second consecutive week of lorazepam use in the first month only. Neither drug could be administered within 8 hours prior to behavioral assessments.
273
99.9 mg/ml; and high therapeutic, /100.0 mg/ml. For lithium, the corresponding serum level ranges were: non-therapeutic, B/0.49 mmol/l; low therapeutic, 0.50 /0.79 mmol/l; medium therapeutic, 0.80 /1.00 mmol/l; high therapeutic, /1.00 mmol/l. Within each defined therapeutic category, Kaplan /Meier Survival and Cox regression analyses [18] were used to determine time to study discontinuation. Separate analyses were performed based on discontinuation for any reason, discontinuation for a protocol-defined manic episode (manic episode accompanied by an MRS score /16, or one requiring hospitalization), or discontinuation for mania or depression (protocol-defined manic or depressive episode requiring the use of an antidepressant or premature discontinuation because of depressive symptoms). All tests were two-tailed; tests of significance have been previously described [6].
Results Demographic and clinical characteristics at baseline A total of 571 patients were enrolled in the openphase study period; of these, 372 (divalproex, N /187; lithium, N /91; placebo, N /94) were subsequently randomized into the double-blind maintenance treatment phase. The treatment groups were well matched, with no significant differences with respect to age, proportion of male and female patients, or race. Mean MRS, DSS, and GAS scores; mean age of first manic episode; and mean age of first depressive episode were all similar between groups. Approximately half of the patients in each treatment group reported /10 prior manic episodes and/or /10 depressive episodes. Serum drug level measurements sufficient for analysis were available for a total of 331 patients (divalproex, N /169; lithium, N /74; placebo, N /88), who constituted the study group for the present study. The stratification by therapeutic category is shown in Table I. With regard to study discontinuation for any reason, Kaplan/Meier survival curves for divalproex and lithium are shown in Figure 1. The time in maintenance treatment before discontinuation for Table I. Patient stratification by therapeutic category.
Analysis Serum drug concentration
Based on a retrospective analysis of individual patients’ serum drug level ranges, randomized patients who received study drug were stratified into four categories (non-therapeutic, low therapeutic, medium therapeutic, and high therapeutic) for each drug treatment. For divalproex, the valproate serum level ranges used to define these categories were: non-therapeutic, B/49.9 mg/ml; low therapeutic, 50.0 /74.9 mg/ml; medium therapeutic, 75.0 /
Therapeutic category
Divalproex, range (mg/ml)
Lithium, range (mmol/l)
Non-therapeutic B/49.9; N /11 B/0.49; N /5 Low therapeutic 50.0 /74.9; N /45 0.50 /0.79; N /16 0.80 /1.00; N /37 Medium 75.0 /99.9; N /85 therapeutic High Therapeutic /100.0; N /28 /1.00; N /16 N / number of patients. Placebo: N /88.
274
P. E. Keck et al. A
100% 80%
Survival
Placebo
60%
Nontherapeutic
*
Low therapeutic
40%
Medium therapeutic High therapeutic
20% 0% 0
1
2
3
4
5
6
7
8
9
10
11 12
Months of Study
B
100% 80%
Survival
Placebo Nontherapeutic
60%
Low therapeutic
40%
Medium therapeutic High therapeutic
20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study Figure 1. Survival analysis: discontinuation for any reason. Kaplan /Meier survival curves for divalproex-treated patients (A) and lithiumtreated patients (B) who were discontinued from the double-blind treatment period of the study for any reason. Only divalproex at the Medium Therapeutic concentration range (*) demonstrated a statistically significant difference (P B/0.05) from placebo. *Significant difference from placebo (P B/0.05).
any reason (i.e. a mood episode, or early drop out for any reason: medication intolerance, non-compliance to protocol, moved, lost to follow-up) was significantly greater for the divalproex treatment group in the Medium Therapeutic serum concentration range (valproate concentration 75 /99.9 mg/ml) than for placebo (P B/0.05). Median survival time for the Medium Therapeutic divalproex group was 8 months, compared with 4 months for placebo. Among lithium-treated patients, no lithium serum level treatment group differed significantly from placebo. The number of patients who relapsed to protocoldefined mania was too small to permit Kaplan/ Meier significance analysis. However, in comparing all treatments, divalproex treatment in the Medium Therapeutic category demonstrated the highest proportion of patients (85%) who survived to month 8. There were no indications of numerical superiority for any of the lithium serum level treatment groups. Kaplan /Meier survival curves for divalproex and lithium with regard to study discontinuation for mania or depression are shown in Figure 2. The only statistically significant difference from placebo observed for divalproex was for the Medium Therapeutic serum concentration range group. Median survival time for this group was 8 months, compared with 3 months for placebo. No lithium serum level group differed from placebo on this analysis. Median
survival times for placebo, lithium, and divalproex groups are summarized in Table II. Discussion This is the first published study that examines response to divalproex maintenance treatment of bipolar disorder as a function of serum drug concentrations. Significant improvement was observed for patients in the divalproex Medium Therapeutic category (serum valproate concentrations 75.0 /99.9 mg/ml) with respect to two outcome measures: discontinuation for any reason, and discontinuation for mania or depression. The number of patients who experienced the third outcome measure, time to protocol-defined mania, was too small to permit adequate statistical analysis, but appears numerically to support the efficacy of this valproate serum concentration range for delaying time to mania. The significant improvement with regard to discontinuation for any reason can be taken as an overall measure of treatment effectiveness, incorporating both elements of efficacy and drug tolerability. This result suggests that for divalproex the Medium Therapeutic range provides optimal effectiveness in preventing relapse while achieving acceptable tolerability during maintenance therapy in a subgroup of recently manic bipolar 1 patients who may not have a severe form of the disorder [6]. Whether such a
Serum valproate and lithium levels in bipolar maintenance therapy
275
A 100%
* 80% Placebo Nontherapeutic Low therapeutic Medium therapeutic High therapeutic
Survival
60% 40% 20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study B
100% 80% Placebo Nontherapeutic Low therapeutic Medium therapeutic High therapeutic
Survival
60% 40% 20% 0% 0
1
2
3
4
5
6
7
8
9
10
11
12
Months of Study
Figure 2. Survival analysis: discontinuation for mania or depression. Kaplan /Meier survival curves for divalproex-treated patients (A) and lithium-treated patients (B) who were discontinued from the double-blind treatment period of the study for mania or depression (protocoldefined mania or depressive episode requiring the use of an antidepressant or premature discontinuation because of depressive symptoms). Only divalproex at the Medium Therapeutic concentration range demonstrated a statistically significant difference (P /0.003) from placebo. *Significant difference from placebo (P /0.003).
serum range would also apply to other subgroups of bipolar patients in maintenance treatment remains to be determined. A comparison of Kaplan /Meier survival curves for discontinuation resulting from protocol-defined mania (not shown) and for discontinuation resulting from mania or depression strongly suggests that relapse into depressive mood episodes is a more significant concern during bipolar maintenance therapy than relapse into mania. An analysis of the original study data with regard to prevention of bipolar depression has recently been published [19]. This study showed that divalproex was more effective than lithium in delaying time to depressive relapse among patients who had responded to Table II. Median survival time: Discontinuation for protocoldefined mania or depression. Treatment group (N )
Placebo (88) Lithium Nontherapeutic (5) Low therapeutic (16) Medium Therapeutic (37) High Therapeutic (16) Divalproex Nontherapeutic (11) Low Therapeutic (45) Medium Therapeutic (85) High Therapeutic (28) NC, not calculable.
Median survival time, months (95% confidence limit) 3.5 (2, 5) 1 4 4 3
(1, (2, (2, (1,
3) 10) 9) 5)
5 5 8 2
(1, (4, (5, (1,
NC) 10) 12) 3)
divalproex during the open-label period, and among those who had previously been hospitalized (suggesting greater illness severity). Divalproex treatment was also associated with less deterioration than lithium with respect to depressive symptoms. In addition, the combination of an SSRI with divalproex was significantly more effective in treating breakthrough depression than an SSRI alone (SSRI/placebo) [19]. Between-group differences with regard to discontinuation for depression alone did not reach statistical significance in this study. However, the significant improvement associated with divalproex in the Medium Therapeutic range with regard to discontinuation for mania or depression, together with the predominance of depression relative to mania as a cause for discontinuation, suggest some efficacy for divalproex at Medium Therapeutic serum drug concentrations in preventing depressive relapse. This study had the advantage of relatively frequent serum level monitoring over a long period of time. Additionally, efforts to obtain serum levels during essentially trough serum level time periods, and analysis of serum levels by a central lab may have strengthened the results. We averaged serum levels over the time of activity in the study. It is possible that this strategy failed to detect outcome differences related to visit-by-visit fluctuations in serum levels. However, our analyses of stability of serum levels within patients suggest that central tendency of
276
P. E. Keck et al.
serum levels was maintained across time for both divalproex and lithium treated patients. Certain limitations in the original study have been addressed in previous publications [6,7,19,20]. The failure to demonstrate significant improvement in the primary outcome measures may have reflected the stringency of the a priori criteria for depressive or manic relapse. This definition in part contributed to a lower than predicted number of relapse events, reducing the power of the statistical analyses [6]. An additional contributing factor may have been the low rate of relapse in the placebo group relative to earlier placebo-controlled double-blind studies of lithium, potentially reducing the effect size. The low rate of placebo relapse may also have been a function of the study duration, as placebo-treated patients in an earlier lithium trial continued to relapse after 12 months [21]. The failure to demonstrate treatment efficacy for any dose of lithium in this study is consistent with the original study, and may also reflect to some extent the low rate of placebo relapse [6]. In addition, the number of subjects in the lithium arm of the study, which was included principally as an active control group, was essentially half of the number in the divalproex group, reducing power. Because the efficacy of lithium in bipolar maintenance treatment continues to be supported by recent trials that were not compromised by earlier methodological failings [5,8,22,23], these results should be interpreted with great caution. However, the results do not tend to indicate that one specific serum lithium range, such as that which is established as effective for treatment of acute mania, is overall more effective than other serum level ranges. Because the linkage of lithium responsiveness in maintenance to serum lithium levels remains a matter of controversy, these results remain of interest in that they would seem to support the case of those authorities who recommend maintenance treatment with lithium at lower than 0.8 mEq/l for some patients. This study was also limited by the relatively small number of patients within the stratified therapeutic categories, reducing the statistical power of the applied analyses. The use of discontinuation for any reason as a measure of tolerability is to some extent compromised by the inclusion of relapserelated discontinuations in this variable. It would be desirable in future studies to provide a separate analysis of discontinuations related to adverse events. Notwithstanding these limitations, this study demonstrates a significant treatment effect for divalproex, when dosed to achieve serum valproate concentrations from 75.0 to 99.9 mg/ml, in the prevention of relapse during maintenance therapy of bipolar disorder I. The novel use of serum valproate concentrations to assess efficacy may help clinicians identify treatment targets for bipolar
maintenance, and establishes a new direction for further research into the effectiveness of divalproex in this challenging therapeutic arena. Key points . Retrospective analysis of 12-month, doubleblind, placebo-controlled trial of divalproex sodium and lithium involving 372 patients (intent-to-treat) . Efficacy measures were Kaplan /Meier survival analyses of time to discontinuation for any reason, time to discontinuation for a protocoldefined manic episode, and time to discontinuation for a manic or depressive episode . Serum valproate levels of 75/99.9 mg/ml were most effective in preventing relapse to mania or depression and discontinuation for any reason. The number of patients who relapsed to protocol-defined mania was too small to permit adequate statistical analysis Statement of interest The authors have no conflict of interest with any commercial or other associations in connection with the submitted article.
References [1] Judd LL, Akiskal HS, Schettler PJ, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry 2003;60:261 /9. [2] Judd LL, Akiskal HS, Schettler PJ, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry 2002;59:530 /7. [3] Practice guideline for the treatment of patients with bipolar disorder (revision). Am J Psychiatry 2002;159:1 /50. [4] Suppes T, Baldessarini RJ, Faedda GL, et al. Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Arch Gen Psychiatry 1991;48:1082 /8. [5] Bowden CL, Calabrese JR, Sachs G, et al. A placebocontrolled 18-month trial of lamotrigine and lithium maintenance treatment in recently manic or hypomanic patients with bipolar I disorder. Arch Gen Psychiatry 2003;60:392 / 400. [6] Bowden CL, Calabrese JR, McElroy SL, et al. A randomized, placebo-controlled 12-month trial of divalproex and lithium in treatment of outpatients with bipolar I disorder. Divalproex Maintenance Study Group. Arch Gen Psychiatry 2000;57:481 /9. [7] Bowden CL, Swann AC, Calabrese JR, et al. Maintenance clinical trials in bipolar disorder: Design implications of the divalproex-lithium-placebo study. Psychopharmacol Bull 1997;33:693 /9. [8] Gelenberg AJ, Kane JM, Keller MB, et al. Comparison of standard and low serum levels of lithium for maintenance treatment of bipolar disorder. New Engl J Med 1989;321: 1489 /93. [9] Hopkins HS, Gelenberg AJ. Serum lithium levels and the outcome of maintenance therapy of bipolar disorder. Bipolar Disorder 2000;2:174 /9. /
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
277
Serum valproate and lithium levels in bipolar maintenance therapy [10] Sproule B. Lithium in bipolar disorder: can drug concentrations predict therapeutic effect? Clin Pharmacokinet 2002; 41:639 /60. [11] Bowden CL, Brugger AM, Swann AC, et al. Efficacy of divalproex vs lithium and placebo in the treatment of mania. The Depakote Mania Study Group. J Am Med Assoc 1994; 271:918 /24. [12] Pope HG Jr, McElroy SL, Keck PE Jr, et al. Valproate in the treatment of acute mania. A placebo-controlled study. Arch Gen Psychiatry 1991;48:62 /8. [13] Calabrese JR, Delucchi GA. Spectrum of efficacy of valproate in 55 patients with rapid-cycling bipolar disorder. Am J Psychiatry 1990;147:431 /4. [14] McElroy SL, Keck PE Jr, Pope HG Jr. Sodium valproate: its use in primary psychiatric disorders. J Clin Psychopharmacol 1987;7:16 /24. [15] Lambert P, Venaud G. Comparative study of valpromide versus lithium in the treatment of affective disorders. Nervure 1992;5:57 /65. [16] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. Revised 3rd ed. Washington, DC: American Psychiatric Association; 1987. [17] Spitzer R, Williams J, Gibbons M, et al. Structured clinical interview for the DSM-III-R. Washington, DC: American Psychiatric Press Inc; 1990. /
/
/
/
/
/
/
/
/
/
/
/
[18] Lawless J. Statistical models and methods for lifetime data. New York: John Wiley & Sons; 1982. [19] Gyulai L, Bowden CL, McElroy SL, et al. Maintenance efficacy of divalproex in the prevention of bipolar depression. Neuropsychopharmacology 2003;28:1374 /82. [20] Baldessarini RJ, Tohen M, Tondo L. Maintenance treatment in bipolar disorder. Arch Gen Psychiatry 2000;57: 490 /2. [21] Prien RF, Caffey EM Jr, Klett CJ. Factors associated with treatment success in lithium carbonate prophylaxis. Report of the Veterans Administration and National Institute of Mental Health collaborative study group. Arch Gen Psychiatry 1974;31:189 /92. [22] Tondo L, Baldessarini RJ, Hennen J, et al. Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders. Am J Psychiatry 1998;155:638 / 45. [23] Berghofer A, Kossmann B, Muller-Oerlinghausen B. Course of illness and pattern of recurrences in patients with affective disorders during long-term lithium prophylaxis: A retrospective analysis over 15 years. Acta Psychiatr Scand 1996; 93:349 /54. /
/
/
/
/
/
/
/
/
/
