In this discussion I focus on Sam's diagnosis, his course of illness, the causes for his lack of response to treatment, his noncompliance, and treatment recommendations.

Sam has been said to be psychotic during his manic, mixed, and major depressive episodes. During his affective psychotic episodes, he is described as "bizarre"—jumping on furniture, snaking along the floor, and crawling like an animal. On several occasions his thoughts have become disordered, and at times his speech is incomprehensible.

The presence of psychotic symptoms actually is quite common with bipolar disorder. Approximately 58 percent of patients with bipolar disorder have psychotic features during a manic episode (1). All forms of psychosis may occur during an affective episode, including mood-incongruent, bizarre, and Schneiderian first-rank symptoms (2,3). Psychotic symptoms that occur during an affective episode may present either as disturbances in the content of thought or as a formal thought disorder, or both. Cross-sectional assessment of the type and severity of psychotic symptoms will not consistently help differentiate whether a patient has schizophrenia or mania. The longitudinal course of illness, family history, and other characteristics are more reliable indicators of a diagnosis of mania. The presence of psychotic symptoms, particularly mood-incongruent psychotic symptoms, appears to predict a worse course of illness (4).

Sam's symptoms also suggest that he has a "depressive syndrome" that occurs between his major affective episodes. During these periods, he is described as isolated and fatigued, rarely initiating or sustaining conversation. The differential diagnosis for this depressive syndrome includes side effects of medications, including neuroleptic dysphoria; depression—that is, residual depressive symptoms—or dysthymic disorder, depending on the temporal sequence with the major depressive episode; negative symptoms of psychosis; functional impairment; and fibromyalgia or chronic fatigue syndrome.

Neuroleptic dysphoria is unlikely in Sam's case because this condition is temporarily associated with the use of neuroleptics and reverses on discontinuation of this class of medications or with the use of atypical antipsychotic drugs. The persistence of Sam's depressive syndrome since the onset of his illness appears to rule out neuroleptic dysphoria as the cause of this syndrome. The possibility that the depressive syndrome represents residual depressive symptoms is likely. However, we do not know the details of his symptoms—for example, whether he experiences guilt, anhedonia, feelings of inadequacy, and so forth. Neither do we know the specifics of the course of his illness (episodic versus sustained) or details of his previous trials of antidepressants. Therefore, it is difficult to discern whether the depressive syndrome consists of and can be attributable to residual depressive symptoms.

It is also possible that Sam is having negative symptoms of psychosis. In between his affective episodes, he is described as being flat and with poverty of speech, and he often does not initiate or follow through with conversations. Some patients with an axis I diagnosis of bipolar disorder and an axis II diagnosis of schizoid personality disorder may be misdiagnosed as having schizoaffective disorder. However, the co-occurrence of these two diagnoses is extremely rare (5).

Sam's recurrent affective psychotic episodes with bizarre behavior, the downhill course of his illness, the lack of a significant family history of affective illness, and the presence of a depressive syndrome in between his major affective episodes raise the possibility that he may in fact have schizoaffective disorder, bipolar type. However, it is difficult to distinguish schizoaffective disorder, bipolar type, from bipolar disorder type I with recurrent affective episodes with psychotic features and a poor course of illness. Deciding whether the correct diagnosis is the former or the latter is more an academic exercise, because both diagnoses generally require treatment with at least an antipsychotic drug and a mood stabilizer.

Before the introduction of atypical antipsychotic drugs for treatment and diagnostic purposes during a depressive syndrome such as Sam's, I might have recommended adequate trials of antidepressants. However, since the introduction of the atypical antipsychotic drugs, some of which appear to confer thymoleptic properties, I may now instead first recommend a trial of an atypical antipsychotic drug before an antidepressant, especially because an antidepressant may precipitate cycling in patients with bipolar disorder. In addition, if the depressive symptoms turn out to be negative symptoms of psychosis, the atypical antipsychotic drug should help, as these agents have been found to be effective for negative symptoms. However, at present Sam is experiencing a mixed episode with psychotic features. My recommendations for the management of this phase of illness are described below.

The diagnosis of fibromyalgia or chronic fatigue syndrome is difficult to make in the presence of a major mental illness, especially when the patient continues to be quite symptomatic. My recommendation at this time would be first to treat his mood disorder.

The second area I address is the downhill course of Sam's illness. Kraepelin recognized the importance of the course of illness in diagnostic classification. Kraepelin (6) referred to dementia praecox, later known as schizophrenia, as an illness with a deteriorating course, in contrast with manic-depressive illness, which was episodic with full recovery between episodes and which was described as having a good prognosis. However, there is mounting evidence that recurrent mood disorders—once considered "good-prognosis diseases"—are, in fact, often very severe and life-threatening illnesses (7). For many patients, the long-term outcome is much less favorable than previously thought, with incomplete interepisode recovery and a progressive decline in overall functioning (7,8). Thus patients with bipolar disorder are likely to be exposed to multiple medications with an increased risk of side effects, and they may have a higher risk of suicide attempts and early mortality (9).

Sam not only has recurrent affective episodes that have not responded satisfactorily to his treatment plan, but he also appears to have experienced a significant deterioration in functioning. In my experience, this is not uncharacteristic; many patients with recurrent mood disorders may eventually develop significant impairment in their functioning even in the absence of major affective syndromes. Our group described this phenomenon among bipolar patients in a recent longitudinal study (7). The reason for impaired functioning even among euthymic bipolar patients is unclear. Some have suggested that it may be the result of residual symptoms or even impaired cognition (10).

There are several issues to address in the third area of my discussion—Sam's lack of response to treatment. Sam has failed to respond to many different medications in various combinations. There are multiple definitions of treatment refractoriness, and it is difficult to determine treatment refractoriness by taking a history. However, it is estimated that up to 60 percent of depressed patients who are referred for an evaluation for lack of response to treatment may have had inadequate trials of antidepressants (11). In my experience, a similarly high rate of patients with bipolar disorder who have been referred to me for consultation for resistance to treatment have not had truly adequate trials.

For example, I was recently asked to consult for a patient who had treatment-resistant bipolar disorder and who had failed to respond during the manic phase of illness to four different antipsychotic drugs—risperidone, olanzapine, quetiapine, and clozapine—as well as to gabapentin, lamotrigine, lithium, valproate, phenytoin, electroconvulsive therapy, and verapamil. It was reported that the patient could not tolerate the sedating effects of clozapine, and the clinician informed me that the clozapine trial was a failure. On closer observation, it soon became apparent that the "failed" clozapine trial was secondary to the sedation. The patient had been started during the manic phase on 50 mg of clozapine a day, instead of a starting dosage of 12.5 mg a day. The patient was also receiving six other medications, most of which were also central nervous system depressants—lithium, valproate, clonazepam, perphenazine, and gabapentin. Thus, depending on the circumstances, a medication regimen may have to be simplified in order to give a particular medication an adequate trial. Only after the regimen is simplified are we able to conclude that the trial was adequate.

It is important to identify and address the causes likely to be associated with lack of response to treatment. They include an incorrect diagnosis; a comorbid psychiatric diagnosis, such as panic disorder, obsessive-compulsive disorder, posttraumatic stress disorder, and substance abuse; a comorbid medical diagnosis; noncompliance; treatment-limiting side effects of medication; inadequate dosages of medications; psychosocial stressors; and the natural course of illness. Most of these issues have been addressed in this discussion.

Sam's diagnosis appears consistent with the case description; there also appears to be no evidence of comorbid psychiatric or medical conditions that would explain the refractory nature of Sam's illness. He has not reported side effects as a major problem. We are left with a few possibilities, namely, noncompliance, which is clearly the case (see below), incorrect dosages of medications, and the natural course of the illness. I have already provided examples of how incorrect dosages and inadequate trials can occur. I discuss this issue further in my treatment recommendations. Finally, it is possible, that Sam has a severe form of mood disorder that will continue to remain refractory to treatment interventions. However, I believe that we still have several options available.

Another point I wish to raise is the importance of recreating past pharmacological trials by way of the life chart methodology (12). Information gathered should include dosages and duration of the trials, combination treatments used, and responses to and side effects of the different medications. I believe that all patients with recurrent mood disorders should have their course of illness charted. Only then can we have a better understanding of how life events, psychotherapies, and somatic treatments positively or negatively affect the course of illness.

The fourth area I wish to discuss is noncompliance. The case summary indicates not only that Sam has had many different trials of medications in a series of different treatment settings but also that he has refused a majority of them and stopped taking his medications soon after discharge. The case summary provides little information about the trials in terms of dosage, duration, response, and side effects. It is also unclear why Sam refuses or stops taking medications. Is it because of ongoing persistent symptoms of his illness or because of other psychodynamic or psychosocial issues? It also appears that during hospitalizations Sam has experienced sufficient improvement to permit him to be discharged but that his response to treatment is short-lived. It is likely that his ongoing symptoms are partly responsible for his lack of insight, which has resulted in noncompliance.

Sam's lack of insight about his diagnosis and about his need for treatment has remained consistently poor since the onset of his illness. Lack of insight is a major problem; when it is present, clinicians hope that it will be limited to the extremes of a patient's affective episode and that insight will improve during the euthymic phase or when symptoms are mild. However, one study suggests that lack of insight among patients with bipolar disorder can be just as severe as among patients with schizophrenia (13). There is no indication that at any time during Sam's illness he has acknowledged that he has a mental illness or that he requires treatment. Unfortunately, if insight does not improve with treatment, Sam's prognosis remains guarded.

I have several recommendations for Sam's treatment. First I wish to comment on his psychosocial treatment. In a majority of cases that are referred to me because of lack of treatment response, it soon becomes clear that treatment efforts are disjointed. There is a rapid turnover of treating clinicians, who are sometimes involved with a patient for short periods and do not have a sound understanding of the treatment goals. Ideally, the same clinicians should be involved for long periods. They should communicate regularly with one another and develop a plan that coordinates all treatment efforts. Clearly, in Sam's case, the treatment efforts appear to be well coordinated, and the teamwork appears to have been a major reason for his continued ability to live outside a hospital.

I would also like to emphasize the importance of having a stable person (treater) in a patient's life, especially a patient with chronic mental illness. Having a stable treater for the long haul is just as important as having a good pharmacological regimen. The hope is that with time, this stable person will help the patient understand his or her diagnosis and need for treatment and provide support during difficult times. It is clear that there has been more than one stable treater in Sam's life.

I have two basic pharmacological treatment recommendations. It is essential to document Sam's previous medication trials with respect to his course of illness by using the life chart methodology (12). In addition, rather than recommending new and untested treatments, I would recommend retrials of previous medications to ensure the adequacy of these trials—unless a particular trial can be clearly documented as a failure.

First, I would recommend a retrial of clozapine. A growing number of open-label studies conducted over the past decade have shown the efficacy of clozapine for some patients with schizoaffective and bipolar disorder who responded inadequately to or were unable to tolerate mood stabilizers or conventional antipsychotic medications (14,15,16). The experts recommend clozapine as the first-line treatment in refractory mania (17).

To help Sam better tolerate this medication, I would recommend starting clozapine during the mixed or manic phase of the illness. Toleration seems to be better when clozapine is started at this phase than when it is started during the depressive phase (14). It may also be necessary to simplify Sam's medication regimen so that he is better able to tolerate the clozapine trial (18). I would recommend a trial of no less than six months, especially because reports indicate that improvement continues beyond the acute phase of treatment (14,19).

Even if Sam does not respond acutely to clozapine, it is possible that he may have less frequent and severe affective episodes in the long term; only by using the life chart methodology will the treating physician be able to recognize this response. It may be worth considering a trial of a depot neuroleptic in combination with clozapine, with the goal of enhancing compliance and preventing rapid decompensation if Sam abruptly discontinues clozapine.

If Sam continues to have recurrent manic episodes despite adequate dosages of clozapine (300 to 600 mg a day), the treating physician should consider combining clozapine with either lithium or divalproex. The latter is usually required if clozapine is prescribed at dosages of more than 550 mg a day because of an increased risk of seizures; clozapine may be prescribed at a dosage of up to 900 mg a day.

If Sam continues to experience recurrent major depressive episodes, a retrial of lamotrigine in combination with the medications suggested above may be useful. Recent data suggest that lithium is more effective in preventing recurrent manic episodes and that lamotrigine is more effective in preventing recurrent major depressive episodes.

Conducting these trials in a more structured and monitored setting to minimize the risk of noncompliance would be preferable. I would hope that a long trial of clozapine—six to 12 months—would help improve Sam's insight enough so that he would want to engage in treatment.

It is very likely that Sam's weight fluctuations over the course of his illness are secondary to his ongoing symptoms. His weight would probably stabilize as his affective and psychotic symptoms improved with the treatment that I have recommended.

A number of other treatment options may be considered. However, they should be used only if it can be established that other treatments have clearly been ineffective. Other medications that have been reported to have thymoleptic properties from which Sam may benefit include ziprasidone (20) and risperidone (21,22). Case reports and series and open-label studies suggest that topiramate, tiagabine, donepezil, nimodipine, and mexiletine are effective in treatment-resistant affective disorders (23,24,25,26).

Dr. Zarate is chief of the mood disorders research unit of the mood and anxiety disorders program at the National Institute of Mental Health, 9000 Rockville Pike, Building 10, Unit 3 West, Room 3s250, Bethesda, Maryland 20892 (e-mail, zaratec@intra.nimh.nih.gov).