Bipolar Disorder

Algorithms for the treatment of bipolar, manic-depressive illness


International Psychopharmacology Algorithm Project Report, 1995
From Psychopharmacology Bulletin 31(3): 457-507, 1995.
Rapporteur: Trisha Supes, M.D., Ph.D.
Members: Joseph R. Calabrese, M.D.
Philip B. Mitchell, M.D.
Peggy J. Pazzaglia, M.D.
William Z. Potter, M.D., Ph.D.
Deborah A. Zarin, M.D.

It has been widely recognized that the availability of lithium has not resolved the question of how best to treat acute-phase illness or manage the maintenance phase of manic-depressive illness. For the 50 to 80 per-cent of patients with "classic" symptoms who respond well to lithium, treatment algorithms for acute episodes are not needed (see Goodwin & Jamison 1990 for a definitive review). However, long-term followup in the last decade has suggested that prophylactic efficacy with lithium monotherapy may be less than 40 percent (Prien & Potter 1990). Moreover, as can be said for virtually all major psychiatric illnesses, few systematically evaluated approaches to treatment exist for patients who fail to respond to an adequate trial of a standard first-line treatment.

Separate algorithms for the acute and maintenance treatment phases of manic-depressive illness were generated in general accord with agreed-upon definitions of the phases of treatment (Prien & Potter 1990). Given the complexity of presentations in manic-depressive illness where distinc-tions are made between bipolar I and bipolar II, as well as between mania (or hypomania), mixed states, and depression, multiple algorithms would be necessary to address the full range of possible symptomatology. For this presentation, the study group limited itself to classic mania or acute depression (Figure 5), maintenance (prophylaxis) treatment (Figure 6), and the acute-phase treatment of bipolar I (Figure 7).

Figure 5

Explaining the Algorithm for the Treatment of Acute Mania (Figure 5)

Line 1. A diagnosis of mania is made based on standardized criteria (i.e., DSM-IV [American Psychiatric Association 1994]). The initial workup must rule out causes of secondary mania, including medical, neurological, organic, or toxic caus-es (Cummings 1986; Das & Khanna 1993; Krauthammer & Klerman 1978). Additionally, comorbidity such as substance abuse, neuropsychiatric factors such as head injury or history of significant neurological illness, or abnormal results in a neurological test such as EEG or magnetic resonance imaging, etc., should be evaluated (Himmelhoch & Garfinkel 1986; Kadrmas & Winokur 1979; Lyketsos et al. 1993; Stoll et al. 1994; Suppes et al. 1994).

Line 2. Lithium remains the general first choice, especially for a first, second, or third episode (evidence level A) (Bowden et al. 1994; Goodwin & Jamison 1990). History of prior poor response, intolerance to lithium, or presentation with a sub-type of bipolar disorder (e.g., rapid cycling or mixed mania) could justify carbamazepine (evidence level B) or valproate (evidence level A) (Bowden et al. 1994; Freeman et al. 1992; McElroy et al. 1992; Pope et al. 1991; Post 1990; Post et al. 1986; Prien et al. 1988; Secunda et al. 1987).

Line 3. Intramuscular or oral neuroleptics and/or benzodiazepines are considered adjuncts; their use may be necessary for behavioral control until marked improvement is achieved (evidence level B) (Busch et al. 1989; Garza-Trevino et al. 1989). Recent studies suggest that high-dose benzodiazepines alone may be adequate for the management of acute mania (Chouinard 1988; Lenox et al. 1992; Modell et al. 1985). Adjuncts should be used as needed. For lithium, blood level should be at or above 0.8 mEq/L (evidence level A) for at least 2 to 3 weeks (evidence level C) (Gelenberg et al. 1989). For carbamazepine and valproate, efficacy is reported at blood levels within the range used for treatment of epilepsy (evidence levels A and B) (Clothier et al. 1991; Simhandl et al. 1993). Recent evidence from a placebo-controlled trial shows that val-proate had efficacy equal to lithium in the first 3 weeks of treatment for acute mania but had fewer side effects (Bowden et al. 1994).

Line 4. Although a 3-week trial of initial treatment(s) is recommended, a 2-week trial is needed to support a change if clini-cal condition neither worsens nor improves. In all cases and at all levels of the algorithm, compliance should be evaluated and blood levels checked regularly when symptoms persist.

Line 5. In the case of no improvement in symptoms, two options are recommended. In the case of partial efficacy, two other options are recommended. In all cases and at all levels of the algorithm, once efficacy is established, the patient moves on to the maintenance (prophylaxis) phase algorithm.

Line 6. The treatment decision will be made based on the patient's history of response and drug tolerance, and on patient -- physician choice.

Line 7. In the case of no efficacy, it is recommended that ECT be administered 3 times per week (evidence level A) (Mukherjee et al. 1994; Small et al. 1988), or that a second mood-stabilizing agent be added to ongoing treatment (e.g., an anticonvulsant added to lithium or vice versa) (evidence level C) (Desai et al. 1987; Keck et al. 1992; Kramlinger & Post 1989). In the case of partial efficacy, it is recommended that either a second mood-stabilizing agent be added or the dose of the agent already initiated be increased to the maximum tolerated level for 2 weeks (evidence levels B and C) (Keller et al. 1992).

Line 8. Although a 3-week trial of initial treatment(s) is recommended, a 2-week trial is needed to support a change if clini-cal condition neither worsens nor improves. Compliance should be evaluated and blood levels checked regularly when symptoms persist. For ECT, six bilateral treatments are recommended before efficacy is evaluated.

Line 9. A 2-week assessment period is recommended. Because of the lack of controlled studies, it is difficult to estimate what percentage of patients will respond to each new intervention.

Line 10. The treatment decision will be made based on the patient's history of response and drug tolerability, and patient -- physician choice. Consideration of starting or continuing ECT will be based on the same criteria.

Line 11. Where clinical response is persistently lacking, it is recommended that ECT be initiated or extended, or that multi-ple mood-stabilizing agents be used to control symptoms. In terms of polypharmacy, three options are considered: use lithi-um plus two anticonvulsants (carbamazepine and valproate) (evidence level C); use lithium plus the anticonvulsant not tried (evidence level C); or use the two anticonvulsants alone (evidence level C) (Desai et al. 1987; Keck et al. 1992; Ketter et al. 1992; Kramlinger & Post 1989). Where treatment response is partial, the recommendations are as above (refer to Line 7).

Line 12. Although a 3-week trial of initial treatment(s) is recommended, a 2-week trial is needed to support a change if clin-ical condition neither worsens nor improves. Compliance should be evaluated and blood levels checked regularly when symptoms persist. For ECT, six bilateral treatments are recommended before efficacy is evaluated.

Line 13. Most bipolar patients will have achieved some degree of mood stabilization and some decrease in other symptoms associated with acute mania by this time. However, a minority of patients may need additional treatment.

Line 14. The treatment of choice will be based on the most current research findings available.

Line 15. When ECT is the treatment choice (Line 11), but only partial efficacy is achieved, trials of lithium plus two anti-convulsants or of the anticonvulsants alone are options to consider before moving to the options proposed in Line 15. All of the following treatments are still experimental, but efficacy has been shown in some patients who have been unresponsive to aggressive treatment with other agents. Calcium channel blockers, especially verapramil, have been found to be effective in some cases (for review, see Dubovsky 1993). Recent open studies with another calcium channel blocker, nimodipine, suggest that this agent may also help some patients (evidence level C) (Post 1990). Clozapine, often in combination with other agents, decreases mood lability and symptoms in treatment-resistant bipolar patients (evidence level C) (McElroy et al. 1991; Suppes et al. 1992, 1994; Zarate et al. 1995). High-dose thyroid treatment has been effective in some treatment-resistant patients with rapid cycling disorder (evidence level C) (Bauer & Wybrow 1990).

Figure 6

Explaining the Algorithm for the Maintenance Phase of Bipolar Disorder (Figure 6)

Line 1. A diagnosis of bipolar I is established, and while the algorithm for acute mania is being followed, alternative diag-noses and causes of symptomatology are being ruled out. The nature of bipolar disorder is chronic and recurring. In many cases, monotherapy may prove inadequate for long-term prophylaxis of this illness. A number of long-term followup studies found that patients who are responsive to lithium in the acute manic phase and who go on to maintenance treatment develop a recurrence of symptoms over the following 2 to 10 years (Dickson & Kendall 1986; Harrow et al. 1990; Maj et al. 1989; Page et al. 1987; Vestergaard & Schou 1988). There are no substantive data on maintenance therapy with the anticonvul-sants, but preliminary data support the finding that long-term prophylaxis is probably limited with monotherapy.

Line 2. This reflects entrance into the maintenance phase algorithm from the acute mania algorithm at whatever point effi-cacy is established for acute symptoms.

Line 3. MONO refers to lithium or valproate/carbamazepine (acute mania algorithm: Figure 5, Line 3). COMBO refers to lithium plus valproate/carbamazepine or valproate plus carbamazepine or lithium plus valproate plus carbamazepine (Line 7 or 11). EXOTIC refers to treatments that either are not well established or have been shown to help in a small number of patients -- for example, calcium channel blockers, clozapine, or thyroid medications (Line 15). Figure 5 evidence is levels A and B for lithium maintenance treatment (see Goodwin & Jamison 1990 for review; see below also), levels B and C for the use of anticonvulsants in maintenance treatment (Coxhead et al. 1992; Lusznat et al. 1988; McElroy et al. 1991; Post 1990), and level C for all exotic treatments, such as clozapine, high-dose thyroid, etc. (see Figure 5 for reference).

Line 4. If symptoms recur, treatment is modified.

Line 5. An increase (or change) of symptoms suggesting a prodromal or a new manic episode returns the patient to the acute mania algorithm (Figure 5). For partial or no efficacy (i.e., the recurrence of symptoms), the clinician returns to Line 11 or to the line and treatment where a decrease of symptoms was achieved on the algorithm.

Lines 6 & 7. This indicates the treatment that was efficacious in resolving the acute mania episode. Because the medication regime was probably simplified during the maintenance phase, a return to the acute mania algorithm may mean an increase in dose or the addition of another medication.

Figure 7

Explaining the Algorithm for the Treatment of Acute Bipolar Depression (Figure 7)

Line 1. A diagnosis of bipolar I is established, and alternative diagnoses and causes of bipolar symptomatology are ruled out. Entrance into the acute depression algorithm occurs from maintenance-phase treatment. It is assumed that the patient is entering on medications that were adequate to sustain maintenance phase treatment. The case of a patient who is recovering from an acute mania episode and develops a significant depression was not considered.

Line 2. The choice of treatment is based on current medications and history of treatment.

Line 3. Medical causes or contributing factors, including subclinical hypothyroidism, are ruled out. The first step is to increase the lithium level if it is found to be below 0.6 mEq/L (evidence level B) (Gelenberg et al. 1989). Alternatives are to add an antidepressant to prior mood-stabilizing treatment, whether that consisted of valproate, carbamazepine, or lithium (evidence level B) (see Goodwin & Jamison 1990 for review).

Lines 4-6. Whereas the work group was successful in developing an acute mania and maintenance phase algorithm, it was unable to reach a consensus on appropriate treatment of acute bipolar depression following these first steps. The difficulty reaching consensus probably reflects the limited knowledge base that is currently available on the appropriate treatment management of patients with acute bipolar depression (Zornberg & Pope 1993).


Reprinted with permission.

to top


Internet Mental Health (www.mentalhealth.com) copyright 1995-2011 by Phillip W. Long, M.D.