Treatments to Control Cycling
As noted earlier, for about 90% of bipolar disorders, recurrence rather than persistence of acute episodes is the primary problem in psychopharmacological management. During the past two decades there has been a nearly complete absence of controlled data specific to the maintenance phase. The treatment algorithm presented in the following subsections (see

Figure 45-2) draws guidance from two generally trusted but unproven clinical principles: 1) agents useful for treatment of acute mania tend to be useful maintenance treatments, and 2) treatments effective for rapid cycling appear to be useful maintenance treatments for patients with non-rapid-cycling bipolar disorder.

Lithium
Lithium is widely regarded as the treatment of choice for bipolar illness. Placebo-controlled data from studies conducted in the 1960s and 1970s (Baastrup et al. 1970; Coppen et al. 1971; Cundall et al. 1972; Dunner et al. 1976; Fieve et al. 1976; Melia 1970; Prien et al. 1974) led to the expectation that lithium alone is effective therapy for all but 20%-40% of patients with bipolar disorder (Prien and Potter 1990). Widely accepted and widely quoted, this myth leads clinicians to expect that 60%-80% of lithium-treated patients will achieve full remission.

Unfortunately, data accumulated over the last 20 years suggest that this expectation is often unfulfilled. After lithium use became widespread in Britain, readmission rates for patients with bipolar disorder were observed to increase (Symonds and Williams 1981). Retrospective studies of patients with bipolar disorder given lithium or other treatments in open practice often failed to demonstrate any significant prophylactic benefit of lithium (Harrow et al. 1990; Mander 1986; Markar and Mander 1989). Typical of these reports, manic inpatients who were prescribed lithium at the time of discharge had approximately the same rate (30%) of early relapse (i.e., less than 3 months from hospital discharge) as those who received no lithium (28%) (Mander 1986). In an outpatient study of 41 individuals with bipolar disorder who were given lithium prophylaxis and compared with a group treated without lithium, no significant difference in relapse rate between the two groups was found (Markar and Mander 1989). Over a 1.7-year naturalistic follow-up, recurrence of mania was reported in 40% of those receiving lithium prophylaxis versus 42% receiving other treatments (Harrow et al. 1990). Poor psychosocial outcome occurred regardless of whether lithium was administered.

Although these studies were open retrospective trials in which it is likely that less intensive management and noncompliance diminished the protective benefit of lithium maintenance, rigorous double-blind maintenance studies also report disappointing results. In a study of lithium prophylaxis (Gelenberg et al. 1989), patients with three or more prior episodes had a 67% recurrence rate, regardless of whether they were assigned to low or standard lithium maintenance. The overall recurrence rate in those with fewer than three prior episodes was significantly lower (34%), suggesting that lithium is less effective for patients after they experience numerous episodes. This finding is consistent with the observations that lithium is less effective for rapid-cycling bipolar disorder (Dunner and Fieve 1974; Secunda et al. 1985).

Incomplete response is a common occurrence in treatment of bipolar illness. Although few patients enjoy permanent cessation of mood episodes, most patients treated with lithium and probably other prophylactic agents (anticonvulsants) can be expected to derive important clinical benefits, such as improvement in severity, duration, or frequency of their acute episodes. Such relative benefits have a critical clinical impact. During a 10-year follow-up of 103 patients treated at a lithium clinic, there were no suicides, and the mortality rate was substantially lower than expected compared with an age- and sex-matched bipolar sample (Coppen et al. 1991).

In open clinical practice, the benefit of lithium monotherapy is modest. Maj et al. (1998) reported follow-up of a clinical sample of 402 patients with bipolar disorder given treatment with lithium. At 5 years, 113 (28%) had discontinued treatment and 44 (11%) were lost to follow-up. Among the 247 still taking lithium over the entire 5-year period, 94 (38%) had no full episodes but only 35 (14%) were euthymic.

Only 4% of the patients in one sample (Sachs et al. 1994b) were both episode-free and treated with lithium alone, which may reflect the treatment-resistant nature of this patient group seen at a specialty clinic. However, many of these patients were initially referred for treatment of their first episode. The percentage of those with bipolar disorder who achieved sustained full remission was quite low, regardless of treatment. If treatment success is defined as sustained remission over 3 years with therapeutic levels of lithium, carbamazepine, or valproate, more than 75% of cases of bipolar disorder could be considered treatment-resistant.

Carbamazepine
The claim of acute antimanic efficacy of carbamazepine is based on 12 studies (total N = 327) (Ballenger and Post 1978; Cookson 1987; Emrich et al. 1985; Grossi et al. 1984; Klein et al. 1984; Lerer et al. 1987; Lusznat et al. 1988; Okuma et al. 1979, 1981, 1989; Placidi et al. 1986; Watkins et al. 1987). These data are supported by 3 additional studies, which demonstrated antimanic efficacy when carbamazepine was given in combination with lithium or antipsychotics (Klein et al. 1984; Kramlinger and Post 1989; Muller and Stoll 1984).

With respect to prophylaxis, five studies are available showing an overall response rate of 72% in 81 bipolar and schizoaffective patients (Ballenger and Post 1978; Lusznat et al. 1988; Okuma et al. 1981; Placidi et al. 1986; Watkins et al. 1987).

Valproate
The acute antimanic efficacy of valproic acid has been established by 8 controlled studies (6 of them double-blind) (total N = 506) (Brennan et al. 1984; R. Brown 1989; Calabrese and Delucchi 1990; Calabrese et al., unpublished results, 1992; Emrich et al. 1980, 1984; McElroy et al. 1988; Pope et al. 1991; Prasad 1984). These include the 2 largest double-blind, placebo-controlled trials conducted in the past 25 years on bipolar disorder. Among 10 studies of valproate maintenance (which were less rigorously controlled than the above 8 studies), involving a total of 375 patients with bipolar and schizoaffective disorder, the response to valproate prophylaxis, averaged across studies, was 54%.

Clonazepam
There are no double-blind, placebo-controlled studies with clonazepam. In the only double-blind study using a crossover design, the acute antimanic efficacy of clonazepam was compared with that of lithium in 12 patients over a 10-day period (Chouinard et al. 1983). This study revealed a statistically significant advantage for clonazepam in controlling one out of five of the manic symptoms rated. Because lithium is not usually effective during the first 10 days of treatment, clonazepam may not be better than placebo in ameliorating the other four symptoms.

Two studies in which clonazepam was given during maintenance treatment in place of an antipsychotic had contrasting results. In an open study, five of five antipsychotic-refractory patients experienced recurrence of psychotic mania within 12 weeks of switching from an antipsychotic to clonazepam (T. A. Aronson et al. 1989). Although this study suggested a lack of efficacy of clonazepam, no treatment has been shown to be effective for bipolar disorder refractory to antipsychotic treatment.

A similar open design for the treatment of patients with bipolar disorder maintained on lithium and an antipsychotic had very different results (Sachs 1990). This study showed that patients could be safely switched from antipsychotic maintenance to clonazepam maintenance. Furthermore, over a 12-month follow-up period, there was a trend showing a lower incidence of relapse into depression for clonazepam-treated patients and no difference in the rate of recurrence of mania/hypomania. A blind comparison of clonazepam and lorazepam in acute mania showed lorazepam to be superior (Bradwejn et al. 1990). It is difficult to account for the results of this study, in which no response to clonazepam was observed.

At this time, the relative advantages of various benzodiazepines are unclear, and double-blind, controlled data are unavailable. Given the relative safety of benzodiazepines, however, their use as adjuncts to other maintenance therapies appears justified. As one anticycling strategy, clonazepam may be given as an adjunct to other maintenance medication in a single bedtime dose of 0.25-4.0 mg; few patients require more than 1 mg, and the majority are maintained at 0.5 mg per day. Clonazepam may also be used in the acute phase or to control manic roughening. In this case, clonazepam is given with close monitoring 0.5-2 mg every 2-4 hours (up to 8 mg/day) until the acute target symptoms have resolved or the patient is asleep. The dosage required acutely is then usually continued in a divided twice-daily regimen over a period of 1-3 weeks.

Lamotrigine
In a double-blind crossover study (Frye et al. 1999), lamotrigine was revealed to be more effective than both placebo and gabapentin in the treatment of both depression and mood elevation. A double-blind multicenter trial (Calabrese et al. 1999) showed lamotrigine monotherapy (both 50 mg and 200 mg) to be superior to placebo for treatment of bipolar depression. Although double-blind data are not yet available for lamotrigine as treatment for rapid cycling, Bowden et al. (1999a) reported encouraging remission rates when lamotrigine was added to the regimen for rapid cycling in the open phase of a double-blind study. High incidences of rash have been associated with high initial doses, rapid titration, age less than 16 years, and the presence of enzyme inhibitors such as valproate. Approximately 0.1%-0.3% of adults treated with lamotrigine are reported to develop serious dermatological reactions (e.g., Stevens-Johnson syndrome, toxic epidermal necrolysis) Recommendations are to start lamotrigine at a dosage of 25 mg/day and to slowly titrate the dose by 25 mg every 1-2 weeks until reaching 200 mg or the best clinical response. The dosing regimen is reduced by 50% in the presence of concomitant treatment with enzyme inhibitors.

Gabapentin
Two double-blind studies have failed to detect antimanic or antidepressant effect of gabapentin. Nonetheless, in a substantial number of open reports this medication has been reported to be beneficial for both depression (L. T. Young et al. 1999) and mood elevation (Dimond et al. 1996; Ghaemi et al. 1998). The lack of efficacy of gabapentin in controlled trials may not mean that the drug is totally ineffective, but at the very least it suggests that its benefit is below the limit of detection in feasible randomized controlled trials. Gabapentin is initially dosed at 300 mg qhs tid and may be titrated aggressively in many patients. Gabapentin is absorbed by an active transport mechanism. Multiple daily doses are usually required due to the drug’s short half-life and because, as dosage increases, the absorption mechanism becomes saturated and serum levels are no longer proportional to dose (L. T. Young et al. 1999).

Hypermetabolic Thyroid
Thyroid hormone has been used for treatment-resistant bipolar illness since the 1930s, when Gjessing (1938) reported positive response to hypermetabolic doses of thyroid hormone administered to patients with periodic catatonia. The relationship between rapid cycling and hypothyroidism remains the subject of some debate. Studies have shown that hypothyroidism (grade I or grade II) is present in 31%-50% of patients with rapid-cycling versus 0%-2% of patients with non-rapid-cycling bipolar disorder (Bauer and Whybrow 1990; Cowdry et al. 1983). In one study, however, hypothyroidism was found in about 40% of patients with non-rapid-cycling as well as rapid-cycling bipolar disorder (Wehr et al. 1988).

There are no double-blind studies reporting the use of thyroid hormone in patients with bipolar disorder. Open administration of 300-500 μg of T₄ abolished rapid cycling in 5 of 10 patients (Stancer and Persad 1982). In another open trial, high-dose levothyroxine was added to the mood-stabilizing treatment regimen of 11 patients with rapid-cycling bipolar disorder (Bauer and Whybrow 1990). Ten of the 11 (90%) experienced “clear-cut” improvement in the depressive phase, and among the 7 with mania at baseline, 5 (71%) responded. Interestingly, response in these patients was independent of initial thyroid status. At the time of clinical response, however, T₄ and free T₄ were found to be above the upper limit of normal in nearly all subjects, and loss of benefit followed when patients tapered thyroid below supranormal levels (with some patients having tapered thyroid under blind conditions) (Bauer and Whybrow 1990).

Candidates for hypermetabolic thyroid should be screened for possible medical contraindications and started with levothyroxine at 0.1 mg/day. Dosage is gradually titrated upward by increasing the daily dosage by 0.05-0.1 mg every 1-2 weeks as tolerated until the free T₄ is about 150% of the laboratory upper limit or the patient experiences intolerable side effects (e.g., anxiety, agitation, tremor, sweating, or palpitations). Like thyroid augmentation of antidepressant therapy, hypermetabolic thyroid treatment is carried out as an adjunct to a standard mood-stabilizing regimen.

Calcium Channel-Blocking Agents
The rationale for use of calcium channel-blocking agents comes from studies showing abnormalities of calcium metabolism in affective illness (Crammer 1977; Gerner et al. 1977). Cerebrospinal fluid calcium concentrations in patients with bipolar disorder have been found to correlate with mood state (Carman et al. 1979; Jimerson et al. 1979). During the 1980s, trials suggested that calcium channel blockers might be effective alternatives or adjuncts to lithium treatment of acute mania. On the basis of two small double-blind studies, it appears that verapamil has antimanic effects (Dubovsky et al. 1986; Giannini et al. 1984). Unfortunately, verapamil does not appear to be a good choice for patients with treatment-refractory bipolar disorder.

Two cases of patients whose depression was refractory to lithium and carbamazepine later also did not benefit from verapamil (Kennedy et al. 1986). Furthermore, of five verapamil responders in another report, none had undergone a previous failed trial of lithium (Dubovsky et al. 1986).

Dosage of verapamil is 160-480 mg given on a split twice-daily or thrice-daily basis. Verapamil is often used as an alternative treatment rather than as adjunct to other anticycling treatments, largely because of drug interactions with lithium and carbamazepine.

The currently available data suggest that the main use of verapamil will be in patients intolerant of lithium or anticonvulsants. Other calcium channel blockers are structurally different and exert effects on intracellular calcium through different mechanisms. It remains to be seen if other calcium channel-blocking agents such as diltiazem, nifedipine, or nimodipine have therapeutic value for treatment-refractory bipolar disorder.

Antipsychotics
Despite widespread recognition among clinicians of the potential risks associated with long-term use of antipsychotic, patients often receive antipsychotic maintenance treatment (Sachs 1990). Given the frequency of this practice, it is surprising that there are no studies reporting any advantage for antipsychotic maintenance in bipolar disorder. Recurrence of mania or psychotic symptoms is associated with tapering of antipsychotic and provides a clinical rationale for this treatment. However, the high frequency of maintenance antipsychotic use may reflect an underrecognition of other treatments for acute mania.

The only studies of antipsychotic maintenance in bipolar disorder suggest that this approach is seldom beneficial. A 2-year double-blind crossover study compared the antipsychotic flupenthixol with placebo as adjuncts to lithium maintenance. Patients receiving flupenthixol experienced more episodes of depression than did placebo-treated patients (Esparon et al. 1986). In another study, no significant difference in outcome was observed between patients with bipolar disorder who continued to take lithium and an antipsychotic versus those who were openly switched to lithium and clonazepam, and a higher incidence of depressive relapse was observed in the patients treated with an antipsychotic (Sachs 1990).

Atypical Antipsychotics (Clozapine, Olanzapine, Risperidone, Quetiapine)
The atypical antipsychotic agents appear to offer better results than the older antipsychotics (Suppes et al. 1999).

Clozapine  In case reports and open series, bipolar disorder refractory to antipsychotics, anticonvulsants, and lithium are reported to derive substantial improvement that persisted in the maintenance phase. Clozapine was reported to be beneficial for seven of seven patients with mixed episodes (Suppes et al. 1992). Treatment was continued for 3-5 years, during which time patients made substantial improvement in psychosocial functioning. Case reports (Frakenburg 1993) describe a positive response during maintenance treatment of bipolar disorder previously refractory to lithium, anticonvulsants, and typical antipsychotic drugs. Dosages of clozapine ranged from 50 to 600 mg/day.
Although they need to be replicated in rigorous double-blind studies, these very encouraging results justify the use of clozapine for treatment-refractory bipolar disorder. Clozapine therapy carries significant risks and limitations, including the need for weekly monitoring of white blood cell counts; the risk of agranulocytosis, seizures, and drug interactions; and high cost. It is reasonable to reserve clozapine for patients unresponsive to other therapies in the algorithm while waiting for safer drugs with a similar profile.

Olanzapine  Open reports of the efficacy of olanzapine for mania and refractory rapid cycling are now buttressed by two placebo-controlled, double-blind studies that showed olanzapine to be more effective than placebo for the treatment of manic and mixed episodes (Tohen and Zarate 1998; Tohen et al. 1999). The antimanic benefit of olanzapine for nonpsychotic and psychotic patients was similar.
Benefit from olanzapine has been reported at daily dosages ranging from 2.5 to 30 mg. Olanzapine has been associated with substantial weight gain, but with early intervention this can often be controlled by diet and exercise.

Risperidone  Risperidone has also been reported to be effective for mania and refractory rapid cycling. A small (N = 45) double-blind study (Segal et al. 1998) demonstrated no difference among risperidone, lithium, and haloperidol. A larger double-blind study (Sachs 1999) compared the acute antimanic benefit of combined treatment with mood stabilizer (lithium or valproate) plus risperidone to mood stabilizer plus haloperidol, or mood stabilizer plus placebo. In this study, the combined treatment with risperidone was significantly more effective than mood stabilizer plus placebo and was equally effective as mood stabilizer plus haloperidol.
Risperidone has been reported to be beneficial at daily dosages between 0.5 and 6 mg. Hyperprolactinemia has been reported more frequently with risperidone than with other atypical antipsychotics. It is appropriate to monitor women with menstrual irregularity or other clinical signs of increased prolactin.

Quetiapine  Although there are no controlled reports on this atypical antipsychotic for bipolar illness, case reports and uncontrolled case series suggest quetiapine may also be of value for treatment of mania and rapid cycling. Regular ocular examination is recommended, because the development of cataracts has been observed in beagle dogs. Benefit from quetiapine for patients with bipolar disorder has been reported at daily doses between 25 and 600 mg.

Adrenergic Blockers
The catecholamine hypothesis of affective disorders provides a simple theoretical basis for using adrenergic blocking agents for the treatment of mania. Open data (M. Hardy et al. 1986; Jouvent et al. 1980; Zubenko et al. 1984) suggested that clonidine was an effective alternative or adjunctive treatment for acute mania. Antimanic response appeared within days of reaching an effective dose (0.2-0.6 mg bid). Of two double-blind studies, one showed clonidine to be inferior to lithium for acute mania, and the other showed clonidine to be inferior to verapamil for lithium nonresponsive acute mania. The meaning of these findings is unclear because of the crossover design, the absence of placebo control, and the sample condition being lithium-responsive in one study and lithium-refractory in the other.

Although no maintenance studies have been reported, clonidine may be a useful adjunct to other mood-stabilizing agents. As noted above, clonidine maintenance has appeared to be beneficial for bipolar disorder with comorbid anxiety disorders in a sample of patients at the Massachusetts General Hospital (G. S. Sachs, unpublished findings, November 1995). Therapy with clonidine is most often limited by sedation and hypotension.

Propranolol, a β-adrenergic blocker implicated as a possible cause of depression, has been reported to be helpful in cases of mood elevation (Von Zerssen 1972). Although some acutely manic patients may tolerate the high doses of propranolol necessary for an antimanic effect, the use of propranolol as a maintenance strategy is not likely to be well tolerated.

Other Treatments
Given the desperation associated with uncontrolled bipolar illness, other therapies might be considered on the basis of case reports or theoretical considerations for patients whose disorder is refractory to the previous interventions in the treatment algorithm. Among these therapies are acetazolamide, spironolactone, amiloride, estrogen/progesterone, and phototherapy. The uncertain benefit of these treatments suggests that their use should be limited, but under most circumstances they can be used safely in the treatment of otherwise treatment-resistant bipolar disorder.

Psychosurgery
Only when a patient continues to experience severe episodes despite exhaustive treatment trials as called for by the treatment algorithm should psychosurgery be considered. Positive outcome has been described in a series of nine patients treated with subcaudate stereotactic tractotomy (Lovett and Shaw 1987). Unfortunately, interpretation of these results is limited by the lack of a control group.

Electroconvulsive Therapy
Use of ECT may be appropriate at any time during the course of bipolar illness. ECT is a particularly attractive option for treatment-resistant bipolar disorder and should always be considered as a treatment option rather than held as a treatment of last resort. For some patient groups, including patients who are acutely suicidal, psychotic, medically complicated, or pregnant, ECT may be the treatment of choice. As mentioned earlier, a course of ECT for acute mania or bipolar depression typically consists of 4-12 seizures induced electrically while the patient is under anesthesia. ECT may be an effective treatment for mania, with a remission rate of approximately 80% (Mukherjee et al. 1994). In large open case series, acute antimanic ECT treatment appears to have been superior to lithium or the combination of lithium and an antipsychotic (Black et al. 1987; Thomas and Reddy 1982). Most studies have demonstrated unilateral ECT to be effective for mania. In one study, however, unilateral nondominant ECT resulted in worsening of the mania in six of six patients treated for mania, but all six improved when treatment was changed to bilateral ECT (Small et al. 1985). A high relapse rate is expected if successful treatment with ECT in the acute phase is not followed by continuation and/or maintenance antidepressant or antimanic prophylaxis (McCabe and Norris 1977; Stromgren 1973).