Short- and Long-Term Psychopharmacological Treatment Strategies

Ira D. Glick, M.D., Steven Marder, M.D., David Janowsky, M.D., Trisha Suppes, M.D., and Charles DeBattista, M.D.

In the previous three volumes, there were no chapters on strategies of treatment. Now that it has been convincingly established that psychiatric illness can be defined and treated as effectively as most other medical illnesses, the question is no longer whether psychotropic medications should be prescribed. In our opinion, the clinical issue is when should medication be prescribed, when is it effective, and under what circumstances should the dosage be reduced or the medication be discontinued for each specific disorder. This strategic issue bears on both short-term acute and continuation phases, as well as long-term maintenance phases.

Why has this set of very different treatment questions arisen in the past decade? First, many active episodes of Axis I disorders have been found to have relatively long durations (e.g., 6 to 12 months for mood disorders and 6 to 18 months for schizophrenia) and many, if not most, Axis I disorders evolve into a chronic vulnerability for repeated episodes or assume a smoldering, subsyndromal course. In contrast, conventional wisdom in the 1960s stated that most episodes of depression were short-lived, self-limited, and would not recur if psychotherapeutic or psychoanalytic treatment was successful—or even if no treatment was given. Second, some Axis I disorders have been found to recur more frequently and become more severe over time, in part because neurobiological mechanisms may induce potentiation or sensitization of underlying neural deficits (e.g., in bipolar disorder). Third, some data suggest that, for the prevention or recurrence phase, once a patient goes off medication, if recurrence occurs, a previously effective medication may no longer be as effective (e.g., as in the case of lithium treatment of bipolar disorder) [40]. Finally, there may be a direct correlation between adequate (mean dose and duration of treatment trial) psychopharmacological treatment and outcome; for many, but not all patients, the probability of poor outcome can be greatly diminished by adequate treatment. Furthermore, good treatment outcome benefits not only the patients but also their families and society in general.

We review here the general rationale and guidelines for short- and long-term psychopharmacological treatment strategies and specifically discuss three important Axis I disorders by way of illustrating current models: schizophrenia, mood, and anxiety disorders. We synthesize some of the best current controlled and clinical data to integrate information about a disorder and its course (e.g., schizophrenia), treatment (e.g., neuroleptics), side effects (e.g., tardive dyskinesia), family variables (e.g., expressed emotion), and societal attitudes (e.g., stigmatization of mental illness) as these factors bear on treatment strategies. We do not cover issues more properly associated with "treatment" (e.g., specific drug selection or the detailed biological deficits of disorders), as they are reviewed in the separate chapters on schizophrenia, mood disorder, and anxiety disorder. Rather, our focus is on strategies of treatment, as such (see also Pharmacological Treatment of Obesity and Disordered Sleep: Developmental and Biopsychosocial Perspectives on the Diagnosis and Treatment of Persistent Insomnia, this volume).

*We are indebted to David Braff, M.D. for contributions to an earlier version, and to Ellen Frank, PhD. for critical review.

Regardless of the disorder, after the acute episode, a major clinical task is not just to reduce symptoms and prevent relapse, but also to raise the level of social and vocational functioning, to approximate as closely as possible the pre-episode levels. A concept that is generally unappreciated is that in some cases and in certain disorders (e.g., schizophrenia and manic-depressive illness), this may take as long as 18, 24, or even 36 months. Furthermore, it is unclear whether the best treatment or treatment strategy is drugs alone or drugs plus family or individual interventions. It is also not well accepted by treating clinicians, patients, or their families, that each episode of a disorder may lead to lower future social, vocational, and symptomatic functioning (47). For example, each episode of schizophrenia can lower function by as much as 50%, and recently the persistent dysfunction with major depressive disorder and bipolar disorder has been documented.

Even among some mental health professionals, the concept of long-term medication strategies may seem extreme and evokes concerns about giving patients and families pessimistic messages about progress and about the potential for negative side-effects. Because these concerns can be ameliorated with educational strategies, we base our recommendations regarding length of treatment solely on the follow-up literature and our knowledge of the natural course of these disorders. There is a wealth of experimental and naturalistic data suggesting that, in a number of disorders, cognitive impairment is enduring, even after an episode resolves. Furthermore, cognitive deficits may worsen with time as suggested by Altshuler for bipolar disorder (2), by Waddington et al. for schizophrenia disorder (56), and by Faravelli and Albanesi for anxiety disorders (15). More broadly, untreated psychotic symptoms have a detrimental effect on subsequent course and outcome (59). The point is that "these mechanisms underlying the expression of psychosis (as distinct from its neurodevelopmental origin) may contribute to apparent progression of the illness" (50).

Needless to say, the challenge in schizophrenia, mood, and anxiety disorders is to identify which patients do not need longer term medication. Unfortunately, there are few data on this question, in part, because we lack definitive data on the natural history, etiology, and underlying neurobiological deficits of the disorders. Thus, our recommendation is to discuss alternatives with the patient and with the family (see below). In general, the clinician should explain that the earlier the onset and more chronic the illness is, the more likely is the possibility of relapse, and that these factors should weigh heavily in deciding whether to continue maintenance treatment.

As a general principle, the most important distinction is to establish whether a patient has a recurrent or chronic fixed condition. The clinician must also decide when an acute episode has resolved and the condition is in full rather than partial remission. If this is the first episode, then the clinician should discuss the probability of recurrence (and its consequences) versus the consequences of maintenance medication. A reasonable general strategy for many disorders is to maintain medication for a first episode patient for 6 months after full remission is achieved. If the condition is established as recurrent, the scale is weighed toward staying on medication for a significantly longer period of time. As a general rule in our experience, patients, and even more strongly their families, when given the choice, prefer the side effects of maintenance medications to the consequences of recurrent bouts of schizophrenia, mood disorders, or anxiety disorders (16). Some recurrent patients prefer to try tapering or stopping drugs after 3 to 5 years of treatment-associated remission, whereas others prefer the prophylaxis of maintenance therapy and will not risk a discontinuation strategy with its associated increased risk of relapse. Increasing evidence in the affective disorders supports the view that after 2 or 3 episodes most patients may need ongoing treatment.

The short- and long-term management of schizophrenia will nearly always include a combination of antipsychotic medications and psychosocial treatments or rehabilitation. Although schizophrenia is frequently a chronic and deteriorating illness, it is important to note that a poor outcome is not inevitable. Patients demonstrate considerable variation in the social, vocational, and symptomatic axes of outcome (47). Also, there is evidence that early treatment with antipsychotics may improve the long-term outcome of the illness. Finally, recent improvements in both pharmacotherapy and psychosocial treatments indicate that the course of the illness can be improved when the optimal treatment is introduced to the right patient at the most appropriate stage of the illness.

In the short term, the primary goal of treatment is to reduce the most dramatic manifestations of acute psychosis. Treatment during the acute phase focuses on the reduction of the most severe psychotic symptoms, including hallucinations, delusions, agitation, and the fragmentation of thought and behavior. Once these symptoms have been reduced, patients often enter a resolving phase during which the individual experiences a reduction in psychotic symptoms, although they remain at increased risk of psychotic relapse if medications are reduced too rapidly or if they are exposed to environmental stresses. The goal at this stage of treatment is to sustain the recent therapeutic gains. The acute phase may last from four to eight weeks, depending upon the responsiveness of symptoms to antipsychotic drugs. The resolving phase may last for up to 6 months. Antipsychotic medication is the mainstay of treatment during these phases; it is unclear if psychosocial treatments are effective when patients are acutely psychotic, and there is some evidence to suggest that overly stressful or intrusive therapies can result in a worsening of symptoms during these phases.

Long-term strategies can be implemented when patients enter a stable or maintenance phase. Psychotic symptoms are often minimal, and the goals of treatment include relapse prevention and rehabilitation. Although the clinician attempts to minimize symptoms and the risk of relapse, it is also important to minimize long-term neurotoxicity, especially tardive dyskinesia. Other important issues during maintenance treatment include the management of negative and cognitive symptoms, which may have a greater impact on the long-term social and vocational prognosis in schizophrenia than psychotic symptoms. Compliance is also an important issue during long-term treatment, particularly when antipsychotics are associated with discomforting side effects.

Despite the differences in these stages, the art of treatment often involves considerations that involve more than one stage. For example, if compliance during the maintenance stage is likely to be a problem, the clinician may choose to treat the patient with oral haloperidol or fluphenazine and to convert the patient to the depot form of the drug at a relatively early stage. Also, if treatment during the acute phase is overly aggressive and results in uncomfortable side effects, patients may become uncooperative during later stages of treatment. Therefore, clinicians may consider treating patients with a low dose of a conventional antipsychotic or a newer antipsychotic if there is concern that even mild side effects will result in a patient's unwillingness to take medications after hospital discharge.

In developing a short-term strategy, the clinician will decide whether the patient will be treated with an antipsychotic, how treatment will be explained to the patient, which drug will be selected and through which route will it be administered, the dose of that agent, and the setting in which treatment will take place. Nearly every patient with an acute psychotic episode will require drug treatment. Although occasional individuals have demonstrated that their episodes are mild and relatively brief, these individuals are the exception. Moreover, as mentioned previously, there is evidence suggesting that delays in initiating treatment with antipsychotics may have long-lasting effects in worsening psychotic episodes and social adjustment (58). In other words, even if a patient eventually recovers without drugs, it is possible that the amount of time spent in a psychotic state may be related to a worse long-term outcome.

Maintaining cooperativeness and trust is an important component of a short-term treatment strategy. The symptoms of acute schizophrenia may include overt suspiciousness or a tendency to misinterpret the intentions of the treatment team. This, plus the cognitive defects, places an additional burden on the clinician that may not be present during the treatment of mood or anxiety disorders. Thus, a successful strategy must include clear explanations of the rationale for treatment and possible drug side effects. Since family members (or significant others/case managers) may be important allies in assuring cooperativeness, family psychoeducation programs are by necessity mandatory in both short- and long-term treatment.

The selection of an antipsychotic has become more complex as newer agents have been introduced. Prior to 1990, all antipsychotics were equally effective and differed mostly in their side effect profiles. Newer drugs (including risperidone, olanzapine, and quetiapine) have milder side effects and, in some circumstances, may be more effective (34). However, these drugs are only available as oral agents and they are much more expensive. The choice of a drug can be made most easily when a patient has a history of a positive response and minimal side effects during treatment with a previous, specific, antipsychotic drug regimen. On the other hand, if previous treatment episodes were associated with discomforting side effects, particularly extrapyramidal symptoms (EPS), clinicians should consider either prescribing a conventional antipsychotic at a lower dose or starting with a newer antipsychotic. If the patient has a history of treatment-resistant positive symptoms with one or more conventional antipsychotics, that individual is likely to be resistant to other conventional drugs (27). As a result, patients with a history of treatment refractoriness should probably be treated with a newer antipsychotic. At this point, there is data suggesting that treatment-refractory patients may respond to clozapine (23) or risperidone (6). The other new antipsychotics have not been studied in this population.

Short-term non-compliance and outright drug refusal are common in individuals with schizophrenia. One of the important determinants of non-compliance is a patient's subjective response to an antipsychotic which, in turn, is related to side effects, particularly EPS (55). This suggests that one of the most important strategies for assuring compliance is minimizing EPS. In younger patients who are vulnerable to dystonias, this may mean prescribing antiparkinson medications along with the first doses of a high-potency antipsychotic. In nearly all cases, patients should be treated with a moderate dose of a conventional drug or with a newer antipsychotic. This strategy for minimizing EPS also suggests that agitation—a common component of severe psychosis—should be managed with adjunctive benzodiazepines rather than higher doses of antipsychotics.

Patients treated with an antipsychotic will usually demonstrate a lessening of acute, positive, psychotic symptoms in a few days to three or four weeks. Some, however, may take up to 12 weeks even when they receive an adequate dose. If a patient fails to respond after four to six weeks of treatment, the clinician should consider ordering a plasma level if the patient is receiving a drug such as haloperidol, fluphenazine, or clozapine where there is sufficient data to interpret the level. If the level is adequate, and the patient has shown no change, the clinician has several options, including raising the dose, augmenting with a benzodiazepine, or changing to another antipsychotic. The relatively high improvement rates on newer antipsychotics—particularly with clozapine— suggest that the most effective option will be to change to a newer drug (38).

The goals of long-term treatment include preventing relapse, supporting psychosocial and rehabilitation efforts, and avoiding drug toxicities such as tardive dyskinesia. There is also the issue of whether recurrent psychotic episodes somehow kindle or facilitate certain neural substrates, as has been posited for mood disorders (40) and thereby increase symptom severity and decrease psychosocial functioning over the long term. This suggests that the long-term objective is to prevent (multiple) episodes that may kindle or potentiate the neural circuit dysfunction basis of schizophrenia and lead to a poorer future course.

Strategies for maintenance treatment should weigh the importance of preventing relapse against the side effects of the antipsychotic medication. This is particularly difficult for the conventional antipsychotics, which often cause EPS at the same doses that are needed for a therapeutic effect. Two major strategies have been studied with conventional drugs. Several groups have utilized an intermittent or targeted drug strategy, which necessitates careful longitudinal monitoring, where medications are withdrawn and then reintroduced at the first symptomatic signs of the onset of psychosis (11). Unfortunately, this strategy as assessed by the Treatment Strategies in Schizophrenia (TSS) collaborative study, results in a high level of relapse (53). Still, this approach may be useful for a minority of patients with relatively mild episodes, good insight, and easily defined prodromal symptoms. An alternative is to use a fixed, low-dose antipsychotic strategy (24, 30). In the long-run, these latter strategies may provide some very important benefits, but they also require a higher level of psychosocial monitoring and assessment on a continuing basis. Therefore, they are also more expensive. Marder and coworkers (32) have proposed a strategy in which patients are treated with a low dose of a long-acting, depot antipsychotic and are supplemented with an oral antipsychotic for prodromal signs of relapse. This augmentation strategy was effective in reducing the risks associated with low-dose treatment.

The dose-reduction strategies previously described may not be relevant for newer antipsychotics, which are associated with a much lower risk of EPS. Unfortunately, there are no controlled, double-blind studies comparing the newer agents with conventional antipsychotic drugs. Nevertheless, experience with the newer antipsychotics, particularly risperidone and olanzapine, indicates that these agents are effective during long-term treatment. If this is proven in long-term studies, it would suggest that these agents can be prescribed at a dose that is most likely to prevent relapse without concern for EPS.

Long-acting, depot antipsychotics are indicated for a substantial number of patients with schizophrenia, including individuals with a history of poor compliance, passive individuals who will accept antipsychotic treatment but will not take them otherwise, and individuals with a history of severe psychotic episodes associated with suicidal or aggressive behaviors.

Psychosocial treatments are most effective when psychotic symptoms have been adequately controlled. The treatment selected will depend upon the patient's and therapist's goals. Many investigators have presented data suggesting that the psychosocial family environment may either potentiate or decrease the level of symptomatology of patients (20). The family that is highly expressive of emotion does not create schizophrenia but certainly can exacerbate symptoms. Specifically, sending a patient with schizophrenia home to an environment filled with conflict, critical comments, and highly charged emotions may act as a stressor. It is also important to recognize that the family does not cause schizophrenia, but that any family with a schizophrenia member may undergo a profound reorganization, and the other family members may feel confusion, guilt, responsibility, and shame. It is thus critically important that the psychopharmacological approach to schizophrenia be combined with family education. Other evidence suggests that social skills training can be effective in improving the social adjustment of patients with schizophrenia (31). Both family education and social skills training are most effective when patients are adequately treated with an antipsychotic. Further, the effects of both treatments tend to wane when they are discontinued, suggesting the importance of prolonged treatment with a need for regular "booster sessions" for individuals who have completed treatment.

Ultimately, when studying the outcome of schizophrenia, ongoing short- and long-term trials are critically important for designing rational drug therapies in combination with psychosocial therapy for schizophrenia patients. The introduction of newer antipsychotics may result in much greater interest in psychosocial interventions. Patients who receive newer agents may be better candidates for psychosocial treatments when treatment with these agents is associated with improvements in negative and cognitive symptoms, as well as reduced side effects. Also, patients who improve on clozapine or other new, a typical drugs may initially appear ready to return to community life. However, these individuals then experience a series of frustrating failures at work, school, or social relationships, which indicates that pharmacologic treatment alone may not be sufficient to prepare them for their new roles. The reader should also refer to Maintenance Drug Treatment for Schizophrenia on maintenance treatment of schizophrenia in this volume for further details.

The primary goal of short-term antidepressant therapy should be to decrease affective symptoms rapidly and to return the patient to his or her original level of functioning—or as near to that as is possible. With respect to attaining this goal, several new directions, innovations, and reappraisals have evolved over the past several years. The first involves a consideration of how long antidepressants should be given acutely before deciding that the effect is inadequate and a change of drugs is needed. Antidepressants as a class require periods of nearly a week to several weeks to begin to exert their effects. Recently, a six- to eight-week trial of full-dose antidepressants significantly increased the yield of responding patients, which represents a change from the earlier consensus that a 3- to 4-week trial is adequate to determine the onset of efficacy (28).

It has become accepted dogma in psychopharmacological circles that the choice of an antidepressant should in large part be made on the basis of side effects, with consideration also given to any prior history of response in the patient or the patient's family. To that end, a drug's anticholinergic, hypotensive, sedating/soporific, or gastrointestinal adverse effects, as well as its overdose toxicity in suicidal patients, should determine the choice. Strategically, the clinician needs to consider the patient's biology and psychology, and this process is not always formally done. Careful attention needs to be paid to the patient's cardiovascular, prostatic, ophthalmologic, and cerebrovascular status, as well as other responses to specific drugs before administering a given drug. This probably necessitates a comprehensive history and physical examination. Conversely, the clinicians should consider the advantages of a given medication's side effects to a specific patient. For example an agitated and insomniac depressed patient may benefit from a sedating antidepressant (44, 52).

Another choice based on side effects is whether to use those drugs that have appeared most recently on the market, and are thus being heavily advertised, or drugs that are "tried and true" and whose advantages and disadvantages have become known over time. With the advent of selective serotonin reuptake inhibitors (SSRIs), such as fluoxetine, sertraline, and paroxetine, and the dopaminergic-activating antidepressant bupropion, a predictable shift in utilization has occurred, especially by primary care physicians. The newer serotonergic agents generally have fewer cardiovascular and anticholinergic side effects, and they are less toxic on overdose than the older, conventional antidepressants. However, the SSRIs are more prone to cause gastrointestinal side effects (which can usually be avoided by starting at a low dose) and to precipitate insomnia, agitation, and psychostimulation. Bupropion may cause relatively more epileptic seizures, although this has not been firmly established.

The treating physician also needs to carefully consider the financial implications of prescribing a particular drug. In spite of differences in side effects, it is important to note that the SSRIs and other new antidepressants show similar efficacies in depressed patients and similar degrees of symptom alleviation as conventional antidepressants. However, the newer drugs usually cost considerably more than generic and even conventional brand-name compounds. Some authorities have suggested using the newer antidepressants as second-line treatments except in rare cases, whereas others have advocated utilizing SSRIs and other new antidepressants initially. In general, however, the practicing community, especially primary care physicians, have shifted their prescribing patterns to the SSRIs, in part because they are more clearly "broad spectrum" than the tricyclic antidepressants (TCAs) and in part because compliance is better with the newer compounds.

Related to the issue of prescribing novel or newer antidepressants for refractory or partially refractory cases of depression is the question of what should be done if a patient is nonresponsive or only partially responsive to a given drug. Such simple measures as raising the dose or measuring blood levels to see if therapeutic norms are being attained or whether the patient is compliant are obvious first strategies, and these measures should be undertaken before switching drugs. In recent years, a variety of augmentation and drug-switching strategies have been developed. For example, the addition of lithium and/or thyroid hormones to an antidepressant regimen for refractory or partially refractory depressed patients has become widely accepted and can increase treatment efficacy. Considerable evidence exists of a genetic component to the efficacy of specific antidepressant drugs, so that if one member of a family has responded well to a specific drug, the chances are high that a positive response will occur in a relative of the index patient. Similarly, a previously effective antidepressant drug will be more likely to be effective when used during subsequent episodes in a specific individual. Finally, monoamine oxidase inhibitors (MAOIs), such as phenelzine and tranylcypromine, appear to have some selective advantage over conventional antidepressants in the elderly and in patients with atypical depression or with variants of bipolar disorder. Use of MAOIs probably should be a second line of treatment (SSRIs being the first line), even in cases of atypical depression. The MAOIs must be used with caution in older patients due to the hypotensive effects.

Lithium and divalproex NA (enteric coated form of valproate) have been approved for treatment of acute mania (18). The results of a recent, placebo-controlled study of lithium vs. valproate in acute mania demonstrated the equivalent three-week efficacy of these medications (7). The accumulating evidence documenting the efficacy of valproate in treating acute mixed and rapid cycling states, combined with the potential to use a loading strategy with this medication, has led to an increase in valproate use (8, 25). Although debate continues concerning the adjunctive use of benzodiazepines and antipsychotics with mood stabilization treatment of acute mania, both are often used for rapid sedative effects. Alternatively, there is growing evidence that the addition of a second mood stabilizer (i.e., carbamazepine or valproate) to lithium is a viable strategy, as opposed to adding an antipsychotic drug (41). In patients non-responsive to lithium, anticonvulsants, or the combination of these medications, clozapine should be considered (9, 59), even when the patient is not psychotic (49). The role of the atypical antipsychotics, including risperidone and olanzapine, in the management of patients with bipolar disorder is not yet clear. Additionally, it appears likely the newer anticonvulsants (e.g., gabapentin and lamotrigine) will be helpful to some bipolar patients. This is an issue currently being studied nationally in controlled trials. Finally, while not often considered as a treatment for acute mania, electroconvulsive therapy (ECT) is a frequently effective treatment option in cases in which nothing else seems to work (36).

Short-term treatment strategies also involve the art, as well as the science, of pharmacotherapy. The treating physician should be thought of as part of the treatment mix. For example, it is important for the prescribing clinician to maintain close and frequent contact with the patient and the patient's family, so that he can monitor side effects and assure compliance. This is especially true when a drug regimen is initiated or when the dose is changed. Contact several times weekly, or even more frequently, either in person or by telephone, is indicated in the early phases of pharmacotherapy. In addition, the fear of stigmatization and the embarrassment often experienced by patients who are prescribed psychotropic drugs may lead the patient to discontinue these agents, a phenomenon that needs to be explored and discussed openly with the patient. There is also a compelling need to discuss potential side effects of the various antidepressant and antimanic drugs, both prospectively and over time, as the patient remains under psychopharmacologic treatment.

By tradition, the relationship of the patient with the therapist has been an insular one. This is especially true in dynamically oriented psychotherapy, which is traditionally a one-to-one situation and usually excludes family members. However, treatment of patients with mood disorders severe enough to require pharmacotherapy often benefits enormously when family members or significant others are involved in the process (19, 20). This is best accomplished by bringing them in for the initial interview as sources of information. This helps to form an alliance with family members as well as the patient. For patients requiring medication, family members should be advised in depth about the nature of the prescribed medication and its side effects. In addition , family members can and should be utilized as helpers in undercutting the patient's resistance to taking medications, making sure medications are taken regularly, and monitoring side effects. Finally, the impact of the patient's illness, as well as the medications and their side effects, on the patient's family need to be considered and handled appropriately.

In any discussion of short-term strategies for utilizing antidepressants, it is important to step back and consider the strategy of not prescribing medication in a given case. In placebo-controlled outpatient trials of antidepressant drugs, approximately 30–40% of patients with major depression responded to the placebo tablet and attentive clinical management with significant alleviation of their symptoms. These patients often, but not always, tend to have episodes that are associated with stress. The episodes have been precipitated by obvious social causes and are less likely to relapse over time. Thus, an important and difficult-to-employ strategy for the prescribing physician is to monitor the patient with social/psychotherapeutic treatment only, to see if symptoms decrease significantly over a period of three to six weeks. This strategy is often difficult for the patient and the prescribing physician, because there is usually a great pressure to "do something." However, used judiciously, in a fairly acute, short-term, non-severe case, this strategy may avoid the costs and side effects of antidepressant use.

Finally, it is important to note that several antidepressant psychotherapies have shown efficacy in the treatment of mild depression (46), and a combination of psychotherapy and pharmacotherapy may be more effective than either treatment alone (5, 42).

In bipolar disorder, the issue of prescribing a medication is clearer for short-term strategies. Any episode of acute mania should receive treatment, and the patient should be evaluated for exogenous causes of the episode, including use of antidepressants, non-psychiatric drugs, or illegal substances (3). This recommendation is based on the evidence of "episodes beget episodes" in bipolar disorder (10) and the low placebo response rate of patients in an acute episode (7). Additionally, the risks to relationships, social functioning, and physical well-being, including the risk of suicide, dictate treatment. The treatment and management of bipolar acute depression has not been well studied (60), The usual practice has been to treat severe or persistent depressive states in bipolar patients (3).

The major goal of short-term treatment of mood disorders is essentially to re-establish mood and functioning to relative normalcy. In contrast, the goal of long-term antidepressant or mood-stabilizing treatment is to maintain the previously depressed, manic, or hypomanic patient at the best level of functioning possible, and, most importantly, to prevent relapse. In the past few years, several studies have demonstrated that the risk of rapid relapse from a given episode of major depression is high if less than four to six months of continued full-dose antidepressant therapy are given following apparent symptomatic remission (13, 28). In addition, it appears from several studies that antidepressants should be continued in full doses during the entire period that a patient is receiving them (17). Following two or three manic episodes, general practice is to continue mood stabilizing medication indefinitely (48). While a small number of bipolar I patients may be free of recurrences over five years, the majority (>90%) will experience a new episode (4). Bipolar patients may be especially susceptible to mood destabilization with abrupt changes in medication (14). If medication changes are needed, a taper and overlap strategy is recommended unless medical necessity dictates abrupt discontinuation. It is possible to decrease the levels of a given mood stabilizer originally prescribed at relatively high doses in order to treat acute mania, but there is little data to provide guidelines. As with all medications, the minimum dose to sustain remission and minimize side effects should be used. Importantly, one of the principals of long-term treatment of bipolar disorder is to aggressively treat residual subsyndromal symptoms including impaired concentration, distractibility, irritability and so on. Recent data suggest that inter-episode dysfunction may be underestimated in affective illness, even following resolution of an acute episode (12, 35).

The latter suggestions for relatively long-term treatment of affective disorders represent a major shift in the strategy for treating affective episodes over time. In the past, the alleviation of a depressive or manic episode was often followed by rapid reduction or termination of medications. Knowledge that the frequency of relapse is high in the short run after an affective episode would indicate a four- to six-month drug treatment phase for virtually all patients started on antidepressant or antimanic drugs. Furthermore, until relatively recently, many clinicians considered it good clinical practice to taper medications once remission had occurred, because the risk of side effects would intuitively appear lower with lower doses of medication. However, tapering of antidepressant drugs leads to higher relapse rates, which suggests that a full-dose strategy should be used in patients with mood disorders in the acute phases and beyond (13, 60).

After a four- to six-month period of active antidepressant (or antimanic) treatment, a decision needs to be made as to how much longer treatment should continue. To this end, the patient's individual history needs to be considered. Stopping medications may be indicated in a patient who has had a complete remission and no previous episodes. In patients with a history of recurrent (i.e., three or more) episodes of depression, mania, or both, there is good evidence that chronic treatment lasting for several years or even longer will decrease partial and complete relapses (51). Also, many patients with only partial remission of symptoms require ongoing treatment. The lethal potential of even one previous suicide attempt should also be considered when weighing the advantages of continuing medication, because the existence and severity of previous attempts by a patient when depressed to no small extent predicts future attempts, as does the disruptiveness of prior episodes. Thus, the earlier strategy of stopping medication in patients with a mood disorder after several months and resuming treatment if and when signs of relapse reappear may be ill advised. The strategy employing chronic, open-ended medication in at least a subgroup of patients appears indicated, as many patients have significant inter-episode subsyndromal depression (20, 26, 35, 51). Also, some patients who were once responsive to either antidepressants for unipolar depression or to lithium for bipolar disorder are no longer responsive when restarted on the same (previously effective) medication (40).

Conversely, however, there are some disadvantages to open-ended antidepressant and antimanic therapy. As described for short-term strategies, psychiatric patients, and especially patients with mood disorders, often feel embarrassed to be taking psychotropic medications. These patients may believe that the consumption of these psychotropic medications is a sign of weakness. They may wish to deny that they have been ill and, in the case of bipolar patients, may even wish to experience the "high" of mania. In addition, there is some epidemiologic evidence that TCAs, given chronically, may facilitate cardiac arrhythmias, leading to significantly increased mortality in patients with preexisting arrhythmias. In addition, antidepressant drugs have significant withdrawal effects if withdrawal is too rapid, and these symptoms may mimic the primary illness. Furthermore, many clinicians have found that much longer tapering periods than previously employed are useful when medication is discontinued, thus lowering the risk of relapse. Finally, the potential for interactions between antidepressants and other drugs increases as the patient ages and is prescribed an increasing number of psychotropic and nonpsychotropic drugs. This may be a drawback to open-ended continuation of therapy. Indeed, it may be considered as an indication for withholding or stopping drug therapy. Nevertheless, at least for the patient with a history of frequent relapses, there are now indications that chronic and continuing maintenance treatment for depression and mania is strategically indicated.

In any case, if an antidepressant drug is withdrawn after a period of continuation therapy, and if symptoms recur, rapid reinstitution of the drug is indicated. If symptoms recur while the patient is still receiving antidepressants or antimanic drugs, a strategy of checking for drug compliance, and use of licit or illicit drugs, the possibility of raising the dosage of drug should be considered, and medical or pharmacodynamic influences should be evaluated.

Several additional strategies exist to deal with manic patients who relapse while being treated prophylactically or while on continuation therapy with lithium. Previously, conventional wisdom suggested the temporary addition of an antipsychotic agent if relapse occurred. The lithium serum level should be raised to acute treatment levels, as well. As an alternative, lithium treatment could be augmented with a mood stabilizer such as carbamazepine or sodium valproate, with ongoing continuation of these drugs as indicated. This may be done in combination with the short-term administration of a benzodiazepine such as clonazepam or lorazepam for sleep effects. The latter strategy avoids the risks of giving an antipsychotic drug with such side effects as development of dystonia, dysphoria, and tardive dyskinesia.

Finally, as with acute mood disorder, there is good reason to involve family members in the treatment of patients on continuing medications. During chronic treatment, visits to the physician occur less frequently, and with time the patient's wish not to use (what he perceives to be) a crutch blends with the fading memory of the pain of the episode to greatly increase the risk of noncompliance. It is at this time that the patient's family needs to be involved, to undercut its own resistance to ongoing medication and to help the patient continue needed therapy.

The issues discussed in this section are similar to those in Chapter 91 and the reader is referred to this chapter for further information.

In this section, we will focus on panic disorder, but also mention agoraphobia and generalized anxiety disorder (GAD). For each, we comment on strategies of when to initiate treatment, which drugs to use (although a full discussion of this issue is outside the scope of this chapter), and when to discontinue treatment, that is, long-term strategies. All of the anxiety disorders can be treated effectively with pharmacologic agents that have a variety of serotonergic and noradrenergic actions. Recent work suggests abnormal dopaminergic function might also be involved in the etiology of anxiety disorders (39). As with mood disorders, the major therapeutic aim is to interrupt kindling and potentiation of these disorders, as discussed earlier in this chapter.

The goal of acute treatment is remission of symptoms and return to normal levels of psychosocial function. As to when to initiate treatment, the first issue is to decide if the anxiety disorder is primary or secondary to another disorder, like chronic alcoholism or depression. If secondary, the primary disorder must be treated first. The next issue is to take a careful history of the nature, frequency, and severity of the attacks. In this context, it should be noted that a very high placebo response rate has been seen clinically and in some panic disorder studies. Perhaps even more so than in depression, the clinician may want to temporize for a few weeks with a patient who has had one or two panic attacks. One attack in four years is quite different than four attacks in one year, but the dramatic impact of even one episode can precipitate a disproportionately strong response and lead to a host of psychological and behavioral sequelae. Once the diagnosis of the subclass of anxiety disorder (panic disorder, GAD, obsessive-compulsive disorder, etc.) is made, the appropriate medication trials can be started.

After the episode has remitted, we recommend continuation of therapy for several months. At that point, the medication should be tapered (slowly, over months) in an attempt to minimize withdrawal and rebound symptoms. Controversy exists about the withdrawal syndrome associated with benzodiazepines. Withdrawal from high doses of benzodiazepines or drugs with short half-lives appears to be more severe (33, 37). As Schatzberg (45) has pointed out, "all available pharmacological treatments may pose some problem with respect to maintenance and discontinuation," specifically MAOIs may cause weight gain, hypertensive crises, hyperpyrexic reactions, and rebound mania. Tricyclic antidepressants may cause weight gain, rebound mania, and increased blood pressure. Benzodiazepines may cause cognitive impairment, ataxia (especially in the elderly), dependence, and withdrawal (45). Even SSRIs are associated with discontinuation symptoms, including dizziness, headaches, and nausea.

It is important to emphasize that patients with chronic anxiety who are effectively treated do not escalate their dose, and that they remain relatively symptom free. Furthermore, if they do become noncompliant, their clinical picture usually worsens.

Agoraphobia, even when it occurs without panic, can be extraordinarily disabling, severely limiting daily activities. In any case, imipramine, alprazolam, and phenelzine have all been found to be effective. General anxiety disorder is most commonly treated with benzodiazepines, but TCAs have also been shown to be effective although slower in onset. Trazodone and buspirone appear to be as effective as benzodiazepines in clinical trials evaluating their short term efficacy in GAD. Post-traumatic stress disorder (PTSD) has not been well studied, but both imipramine and phenelzine will reduce symptoms. SSRIs have been studies in PTSD and appear to have some utility. Medication should be combined with family or individual psychotherapy, as appropriate to each clinical situation.

Many recent studies suggest that all of the anxiety disorders (GAD, panic disorders, OCD, and the phobias) are both very disabling and have chronic courses, rather than being limited to a single episode (15, 43). The Task Force of the Collegium Internationale Neuro-Psychopharmacologicum (CINP) has noted that "it is well documented that the life time prevalence rate for all the anxiety disorders in Europe and North America is approximately 14% of those over the age of 18 years. In addition, such disorders are often associated with other types of mental illness such as depression, alcoholism and drug abuse. Despite the high prevalence of anxiety disorders and sleep disorders which may accompany such conditions, the seriousness of such disorders is often underestimated by the media and by society in general. It has been established, for example, that of those patients with anxiety disorders one-third recover but in the remainder the symptoms persist throughout their lives (50)."

Thus, the cost of anxiety disorders to the individual and to society is considerable and is frequently underestimated. In fact "in a detailed study of over 3000 patients with "pure" anxiety in Sweden, the risk of suicide was found to be as high as in those with depression or other diagnoses that require in-patient care" (1). In part, the high suicide rate may be related to the demoralization that is associated with the day-by-day symptoms in untreated or partially treated patients.

Accordingly, the strategies we advocate are based in part upon data showing that there is severe morbidity in panic disorder, as manifested by depressive symptoms, suicidal behavior, alcohol abuse, and in deterioration of social and vocational functioning (57). A CINP Task Force has noted that "it is well established that the long-term use of anxiolytics and hypnotics can cause dependence in some individuals. However, a significant number of patients suffering from chronic anxiety and insomnia require continuous treatment for periods exceeding the 4 to 6 weeks usually recommended by regulatory authorities. Such long-term treatments may together with other health measures improve the quality of life of such patients and there is no clear evidence that the efficacy of long-term medication diminishes with time or that escalation of the dose may occur in such a way that it may create a clinical or public health problem (50)."

As to which classes of drugs to use for the anxiety disorders, we suggest two possible strategies. The first and more traditional strategy is the use of benzodiazepines. The efficacy and safety of these drugs have been well documented. Most important, these drugs offer large comparative advantages to the classic alternatives (barbiturates, meprobamate, etc.). They are also preferable to no treatment, which—if relapse occurs—increases the likelihood of great individual and social cost (as mentioned above). The second and newer strategy is the use of antidepressants. Benzodiazepines may be good long-term agents for panic disorder and GAD, but they may have more limited utility in OCD, social phobia, and PTSD. In these cases, the SSRIs and other antidepressants appear to be more useful. Venlafaxine and nefazodone have been studied in open-label trials for the treatment of panic disorder, but not all antidepressants appear equally useful for this condition. Bupropion has no demonstrated efficacy that we are aware of for any anxiety disorder. Which medication for which subcategory of anxiety disorders will be discussed in the chapters on treatment of anxiety disorder. Here, in the context of discussing strategies, the relevant issues are that antidepressants are an effective alternative to benzodiazepines. While antidepressants have a longer onset than benzodiazepines, they do not always have greater untoward effects (54). They do, however, have a lower risk of dependence. Until controlled head-to-head comparisons are made, more definitive strategies cannot be suggested. There is sometimes a role for antipsychotics in the management of nonpsychotic anxiety, and these may be as effective as benzodiazepines but may also carry additional risks. In addition, beta blockers can be used to manage social phobia and test anxiety, while antihistamines such as hydroxyzine are indicated for and occasionally helpful for some patients with generalized anxiety who are at risk with benzodiazepines.

As to when to stop therapy, we recommend considering discontinuation after six months of sustained improvement. If no relapse or recurrence is observed, then periodic visits and contact with the patient and the patient's family seem sensible. If symptoms increase, then treatment for another year may be a serviceable guideline. We do not as yet know what percentage of patients will require maintenance treatment, but our guess is that it may be 40% of those who initially present with DSM-IV criteria for panic disorder.

In summary, Schatzberg (45) has delineated the issues well: overall, the risk of anxiety disorders causing morbidity and becoming chronic appears to be much higher than the risk of a treatment being injurious to patients. Given the responsiveness of the disorder to currently available therapies, the overall risk-to-benefit ratio for treating panic disorder patients weighs in favor of treatment.

The following is a summary of some of the new and most important guidelines for pharmacological strategies. They are based on the assumptions that discreet diagnostic entities can be identified, that effective pharmacologic treatments exist, that most of the Axis I and II disorders are life-long or represent life-long vulnerabilities, and that, most importantly, exacerbations and recurrences are often disastrous to patients, their families, and the society in which they live.

In the short term, the goals are to reduce symptoms as rapidly as possible, build an alliance for long-term management, educate the patients and their families about the illness, its treatment, and its course (treated and untreated), and lay the groundwork for as close a return to premorbid function as possible. Effective strategies include the use of an adequate dosage for each medication, adequate duration of use before abandoning the treatment, avoidance of polypharmacy whenever possible, combining medication with psychotherapeutic strategies (almost always, for both patients and their families), and providing systematic (and repetitive) psychoeducation for patients and their families. Almost all acute treatments should continue for at least six months, and with disorders like schizophrenia may take 18 months until symptom remission. Further, we recommend a discrete, six-month period of remission before starting to taper therapeutic medications.

In the long term, that is, during the prevention phase of treatment, we recommend the primary goal of minimizing the risk of relapse, because, for most patients and their families, relapse means marked impairment in social and vocational function, as well as symptom exacerbation. In our experience, many patients and families prefer the consequences of staying on medication for life, rather than the consequences of discontinuing medication. Therefore, we recommend the strategies of (a) staying on medication in an effective dose, (b) frequent monitoring of side effects and life circumstances, and (c) frequent contact with families to maximize compliance and reduce the burden of living with chronic psychiatric illness. We recognize that keeping patients on medication for a lifetime is costly. But we believe the cost is worth it, as there are long-term benefits for the individual, the family, and society. In fact, a Swedish study has shown that initially putting patients on expensive psychotropics saves money in the long-run, because of the lower relapse and rehospitalization rates (29).

In conclusion, we have presented a rationale and a model for the psychopharmacological treatment of acute and chronic psychiatric illness. The data suggests that (a) patients and their families prefer treatment to no treatment because of the devastating effects of recurrence and relapse; (b) if practitioners choose to treat, it must be with an adequate (i.e., effective) dose and an adequate duration of treatment (often life-long). However, there must be a continual search for the minimal dose that yields a positive outcome with maximal efficacy and safety. More data from controlled studies are needed to define more precisely the best treatment strategies over time (i.e., the life course, rather than only for acute episodes). The last four generations of psychopharmacology have seen enormous progress toward developing rational treatment strategies for mental illness based on controlled double-blind studies rather than on theoretical biases. We are confident that this trend will continue.

published 2000