Psychiatric Annals

The Harvard Psychopharmacology Algorithm Project 

Consultant for the Pharmacotherapy of Schizophrenia

David N Osser, MD; Carlos A Zarate, Jr, MD

Abstract

The Consultant for the Pharmacotherapy of Schizophrenia is an Internet-based computerized expert system developed as part of the Harvard Psychopharmacology Algorithm Project. It evolved from the authors' earlier algorithms published in 1988 through 1998,1"6 and the first DOS computer versions of 1991 through 1995.7 It was completely rewritten and updated for the first Internet version, which was completed in 1998, and is now in version 2.0. It may be found at the address www.mhc.com/expert. html. This article provides a summary of the flowcharts and recommendations in this program and is intended to be a companion in printed format for users of the web version. Readers are encouraged to experience a "virtual psychopharmacology consultation" by trying the full program, but must be using an advanced browser such as Microsoft Internet Explorer Version 4.01 to access all the linkages.

The flow of the algorithm is comparable to the series of questions that a psychopharmacology consultant would ask about a patient. Depending on the answers, more questions are asked, or suggestions for further evaluation are made, until there is enough input to enable recommendations. The authors have been seeking to capture the logical process of how a clinician with an expert knowledge base evaluates patient data and selects the best medication for a particular patient (ie, the "art" of prescribing). One vision of the future of health care is that physicians will be able to access Internet "consultants" such as this for decision support at any point of service.8

WHAT IS THE CLINICAL SITUATION OF THE PATIENT?

The entry into the algorithm begins with making a diagnosis of schizophrenia or schizophreniform disorder based on DSM-F/ criteria, and ruling out medical causes of the symptoms. Indeed, throughout the algorithm, the user is cautioned to review or reestablish the diagnosis after each sequential medication trial if the response remains unsatisfactory. Figure 1 displays the flowchart for the initial steps depending on the patient's clinical situation. Prominent in the early assessment of the patient with a psychotic disorder is consideration of the possibility of comorbid substance abuse or dependence, which is present in approximately 50% of patients. This reduces the response to at least some medications and increases the relapse rate.9 If present, "dual diagnosis" treatment will be necessary, but if it is not available, the two disorders will have to be treated separately.10

Next, the Consultant asks which of the following four clinical situations applies best: first onset and initial treatment with an antipsychotic, acute exacerbation or recurrence, agitation requiring parenteral medication, or unsatisfactory recovery following initial treatment.

Although optional in this algorithm, and somewhat contrary to our general preference for minimizing polypharmacy, a trial of one or two adjunctive agents may be considered prior to clozapine for patients with an unsatisfactory response. Unfortunately, these adjunctives are often not particularly helpful,53 so the use of clozapine should not be unnecessarily delayed. The Consultant has much more detail on the rationale and current evidence supporting these various choices than has been included here. Choices are made based on which clinical subsyndrome is present that would comprise the target symptoms for the proposed adjunctive. These are briefly summarized below (Fig. 8).

Manic symptoms? Mood stabilizers: lithium, valproate sodium, carbamazepine, electroconvulsive therapy, or their combination if monotherapy provides partial response.

Impulsive, violent behavior? Lithium, carbamazepine, beta-blockers, or electroconvulsive therapy.

Depression, dysthymia? (occurring in the active phase of the psychosis) If the patient is depressed while taking a neuroleptic, switching to a new-generation antipsychotic is preferred.54,55 Antidepressants are of questionable benefit.56

Secondary ("post-psychotic") depression? Tricyclics and possibly other antidepressants.57

Residual negative symptoms? Be sure secondary negative symptoms are ruled…

The Consultant for the Pharmacotherapy of Schizophrenia is an Internet-based computerized expert system developed as part of the Harvard Psychopharmacology Algorithm Project. It evolved from the authors' earlier algorithms published in 1988 through 1998,1"6 and the first DOS computer versions of 1991 through 1995.7 It was completely rewritten and updated for the first Internet version, which was completed in 1998, and is now in version 2.0. It may be found at the address www.mhc.com/expert. html. This article provides a summary of the flowcharts and recommendations in this program and is intended to be a companion in printed format for users of the web version. Readers are encouraged to experience a "virtual psychopharmacology consultation" by trying the full program, but must be using an advanced browser such as Microsoft Internet Explorer Version 4.01 to access all the linkages.

The flow of the algorithm is comparable to the series of questions that a psychopharmacology consultant would ask about a patient. Depending on the answers, more questions are asked, or suggestions for further evaluation are made, until there is enough input to enable recommendations. The authors have been seeking to capture the logical process of how a clinician with an expert knowledge base evaluates patient data and selects the best medication for a particular patient (ie, the "art" of prescribing). One vision of the future of health care is that physicians will be able to access Internet "consultants" such as this for decision support at any point of service.8

WHAT IS THE CLINICAL SITUATION OF THE PATIENT?

The entry into the algorithm begins with making a diagnosis of schizophrenia or schizophreniform disorder based on DSM-F/ criteria, and ruling out medical causes of the symptoms. Indeed, throughout the algorithm, the user is cautioned to review or reestablish the diagnosis after each sequential medication trial if the response remains unsatisfactory. Figure 1 displays the flowchart for the initial steps depending on the patient's clinical situation. Prominent in the early assessment of the patient with a psychotic disorder is consideration of the possibility of comorbid substance abuse or dependence, which is present in approximately 50% of patients. This reduces the response to at least some medications and increases the relapse rate.9 If present, "dual diagnosis" treatment will be necessary, but if it is not available, the two disorders will have to be treated separately.10

Next, the Consultant asks which of the following four clinical situations applies best: first onset and initial treatment with an antipsychotic, acute exacerbation or recurrence, agitation requiring parenteral medication, or unsatisfactory recovery following initial treatment.

Figure 1 . Algorithm for the initial evaluation and treatment of acute episodes.

Figure 1 . Algorithm for the initial evaluation and treatment of acute episodes.

Flist Treatment With an Antipsychotic?

The first-line recommendation is to select one of the new generation of antipsychotic medications,11'33 olanzapine or risperidone.* Other firstline choices that are slightly less favored at this time are quetiapine and ziprasidone (when available): the evidence for efficacy of these in firstonset situations is less robust. Adequate trials of these agents are briefly described below.

The trials should last approximately 8 weeks. They can be much longer if there is a partial response that has not plateaued,13a and shorter when there is absolutely no response or if there is exacerbation of symptoms or intolerance of side effects. Doses may be different in patients who have compromised metabolism or who are particularly susceptible to side effects. In the inpatient setting, physicians have become accustomed to making changes or additions much more rapidly when the response is unsatisfactory, in response to pressure to find ways to shorten the patient's length of stay. The goal of treatment in this setting is limited to achieving sufficient behavioral control to enable transfer to a less expensive setting for continued care. By contrast, the goal of medication use in this algorithm is to help the patient maximize symptomatic and functional recovery from the schizophrenic syndrome. The adequacy of medication trials that occurred during inpatient or other "crisis" care may need reassessment in accordance with algorithm parameters.

The starting and final dose of olanzapine for the initial treatment of nonsmoking, healthy adults is typically 5 mg/d. Smoking increases clearance by 40%, according to the package insert. Most patients who have chronic schizophrenia smoke. This may explain why most patients seem to do best with 10 mg or more, both in the premarketing trials and in subsequent experience. Female gender decreases clearance by 30%, so the 5-mg dose may be particularly applicable to (nonsmoking) women. The maximum dose is 20 mg, and although it has become common practice to use more in treatment-resistant patients, the necessity of this has not been demonstrated. Olanzapine has been reported to take effect more slowly than standard neuroleptics during the first 2 to 3 weeks of treatment14; therefore, if clinicians were to allow more time at lower doses, the end result might be the same with fewer side effects. Weight gain, elevated triglycerides, and abnormal glucose tolerance seem to be side effects of increasing concern.15

Risperidone is usually started at 0.5 to 1.0 mg twice daily for healthy adult patients receiving their first antipsychotic, and increased to 2 to 3 mg/d during the course of a week. The best results are with doses that do not produce any detectable pseudoparkinsonism (ie, generally less than 6 mg/d even for those who are not neuroleptic naive).12 Risperidone may be a better choice for patients at risk for weight gain, or who quickly gain weight on the other medications. Prolactin elevation and related effects on sexual, menstrual, and reproductive functioning are greater concerns with this antipsychotic.158

Quetiapine may be favored when low parkinsonism, akathisia, and prolactin are particularly desired, but it does cause considerable weight gain and some postural dizziness from alphaadrenergk blockade. It is usually begun at 25 mg twice daily and increased by 25 to 50 mg/d until the dose reaches 200 to 300 mg/d. The maximum dose is 750 mg/d.

Second-line choices for treating the first psychotic episode would be the standard, typical neuroleptic drugs: chlorpromazine, fluphenazine hydrochloride, haloperidol, perphenazine, thiothixene hydrochloride, and trifluoperazine hydrochloride. Not every expert agrees that these agents are no longer first-line choices, and it is difficult to exclude the support of pharmaceutical firms as a contributing factor in an expert's interpretation of the evidence, but the dissenting point of view appears to be a small minority.16 Recent evidence of a markedly lower rate of tardive dyskinesia may be the most important consideration.17

All neuroleptics should be started at a low dose (eg, 1 to 2 mg of haloperidol or equivalent) and raised to a dose that produces no more than minimal parkinsonian side effects. If the patient is clearly improving at a dose that does not produce parkinsonism, it is not necessary to raise the dose further However, the clinician should remain observant for the possibility that the good results were due to hospitalization, placebo effect, positive transference, or other nonspecific factors.18 Details on parameters of an adequate trial may be found in this article within the section on the clinical situation of patients with an unsatisfactory level of recovery.

Acute Exacerbation or Recurrence?

The treatment strategy for an acute exacerbation or recurrence depends on how it happened (Fig. 1). If this was a breakthrough of symptoms while the patient was complying with maintenance pharmacotherapy, then perhaps the medication regimen was suboptimal in that it was unable to blunt the impact of the stress that precipitated the relapse. The algorithm addresses this situation in the same way that it considers the clinical situation of the patient with an overall unsatisfactory recovery.

If the relapse occurred after the patient discontinued medication or took it irregularly, then the Consultant wants to know whether the physician supported the plan to reduce the medication or whether the patient decided this on his or her own. The physician might have agreed with the medication withdrawal because the patient had been doing well and it seemed desirable to establish whether the patient still needed the medication. Unfortunately, this plan did not work, and the patient has suffered a relapse. The question now is whether the patient should go back to taking the previous medication.

This decision depends on how well the patient was doing and what side effects were occurring before the medication was reduced. If the patient had made a full recovery (symptomatically and functionally) and the side effects were negligible, it seems reasonable to resume the previous medication, whatever it was. However, if the patient had not achieved a full recovery, had significant or disabling side effects, or both, then the recommendation would be to try a different medication. The selection should be from the list of firstline choices for the first-onset patient. The patient should have a full trial, with the goal of achieving a full symptomatic and functional recovery. If the response remains unsatisfactory, the patient will fall under the category of "unsatisfactory recovery" in Figure 1 and the algorithm is followed from there.

But, what if the history indicated that this relapsing patient had stopped medication without the approval of the physician? The strategy then would depend on whether the patient's uncooperativeness is considered likely or unlikely to recur. If the patient does not have a good alliance with the physician, has no insight into or understanding of his or her diagnosis and the need for medication, and does not live in a structured living situation where external supervision of his or her use of medication is possible, then all signs point to a need for a depot neuroleptic.19 If any one of these three conditions is not present, however, a trial of one of the first-line antipsychotics should still be considered. This is because these medications offer the possibility of a better outcome even when compared with depot neuroleptics,20 and therefore the risk of relapse associated with a medication switch may be acceptable. If the patient lives in, or moves to, a supervised setting, it may be possible to try a new medication with the help of this structure even if the patient has a poor alliance and little insight.19 Psychoeducational counseling is another intervention that may improve alliance and insight, making it possible to consider a trial with one of the new antipsychotics.21

Even if all three predictors of the need for depot medication are present, the algorithm continues (Fig. 1) by asking whether the alliance with the physician is poor because of severe side effects from medication. If yes, it still may be worth switching to one of the better-tolerated oral medications: if successful, the patient's alliance with the physician and compliance may improve. If not, the recommendation is for a depot neuroleptic.

Agitation Requiring Parenteral Medication?

The standard treatment is 2 to 5 mg of haloperidol intramuscularly every hour, up to three doses, supplemented by 2 mg of lorazepam intramuscularly if the patient is not already overmedicated with sedative-hypnotics.22 Higher doses are usually not needed. If the patient requires more than a minimal amount of haloperidol (eg, 2 to 4 mg), prophylactic antiparkinsonian medication is usually recommended.

Other neuroleptics available in intramuscular form and suitable for this use are loxapine, perphenazine, and thiothixene. Another parenteral medication in use at some centers is droperidol, which may have a somewhat lower risk of side effects compared with haloperidol. The usual initial dose is 5 mg intramuscularly or intravenously, although higher doses up to 50 mg intramuscularly or intravenously have been described.23

Olanzapine, risperidone, and ziprasidone have parenteral forms undergoing clinical trials. These appear promising, and they may eventually replace the standard neuroleptics if their side effect profiles prove more benign. Drugs with more alpha blockade such as chlorpromazine and mesoridazine besylate are also available in intramuscular preparations, but are not favored because they carry significant risk of hypotension, especially at the doses required to get the same effect as the higher potency agents.

Unsatisfactory Recovery With Antipsychotic Regimen?

Here, we are referring to a broad group of patients with persisting disability from positive or negative symptoms; impairment in work, social functioning, or both; need for a supervised living situation; lack of insight into the need for treatment of their mental illness; or limited capacity for self-advocacy.

Figure 2. Algorithm for medication toxicity evaluation.

Figure 2. Algorithm for medication toxicity evaluation.

Step One: Rule Out Toxicity From the Current Medication Regimen. The first step is to be sure that the dysfunction is not due to disabling side effects from the present antipsychotic and other medications. Figure 2 shows the flow from "More #1" in Figure 1. Three possible indicators of toxicity are addressed here: anticholinergic effects, oversedation, or a non-parkinsonian tremor. Extrapyramidal effects are evaluated later.

Symptoms of anticholinergic toxicity can indude confusion, disorientation, visual hallucinations, and memory disturbance. Antiparkinsonian agents, tricyclic antidepressants, and some antihistamines such as diphenhydramine, especially when added to chlorpromazine, thioridazine hydrochloride, dozapine, and perhaps olanzapine, can produce problems in this area. Also, some studies raise the possibility that anticholinergic agents may interfere with or reduce the efficacy of antipsychotics on positive symptoms of schizophrenia.24 For these reasons it is best to keep the anticholinergic load as low as possible for patients treated with antipsychotics.

Sedation is another form of toxidty that may lead to complaints of tiredness, inability to read or concentrate, a feeling of heaviness, or slowed thinking or movement. Potentially causative agents in the patient's regimen could perhaps be reduced or eliminated, such as benzodiazepines, trazodone hydrochloride, valproic add, sedating antihistamines such as diphenhydramine, and tertiary amine tricyclic antidepressants. Even selective serotonin reuptake inhibitors can be sedating for some people. Persistent sedation is one of the "disabling side effects" of neuroleptics that may escape the attention of physidans, although it is a major cause of subjective discomfort and inability to function.25 Once possible causative factors have been eliminated, medication has been given at bedtime, and the antipsychotic dose has been reduced if possible, anecdotal experience suggests it is occasionally helpful to try a small dose of caffeine or bupropion hydrochloride. The patient should be reasonably stabilized.

A third indicator of side effect problems is the presence of a tremor.26 If it is a slow (4 to 6 beats per second), regular, rhythmic ("pill-rolling") movement exacerbated at rest and reduced by voluntary movement, it is probably pseudoparkinsonism: this will be systematically addressed below. At this point, we are interested in the more rapid (8 to 12 beats per second), irregular action tremors that increase with voluntary movement. These are common in patients taking psychiatric medications. Some of the causes of this tremor may be associated with brain dysfunction, including impaired cognition. The differential diagnosis includes high doses of lithium, anticonvulsants, or neuroleptics (the latter also cause parkinsonian tremor); hyperthyroidism; electrolyte imbalance; alcohol and other drug withdrawal syndromes; and Parkinson's disease. Treatable causes should be addressed.

Step Two: Evaluate the Adequacy of the Trial of the Current Antipsychotic. Having completed the assessment and management of some possible cUsabling side effects, we proceed to determine whether the current antipsychotic has been given an adequate trial. Recall that this is a patient whose response has been unsatisfactory or who has had a breakthrough of symptoms despite compliance with the prescribed regimen. Once instruction has been given on how to complete an adequate trial, suggestions are made for what to try next, which depend on what has been tried before. There are four algorithm branches that proceed differently, depending on what medication the patient is taking at the present time.

We begin with the algorithms for what to do if the patient is currently taking one of the standard, typical antipsychotics (possibility #1). Figure 3 displays this rather complex pathway.

1. Is the Patient Currently Taking a Standard, Typical Neuroleptic? The six standard, typical neuroleptics are chlorpromazine, fluphenazine hydrochloride, haloperidol, perphenazine, thiothixene, and trifluoperazine hydrochloride. A "typical" neuroleptic is one that produces parkinsonian side effects. These six medications are considered essentially similar in efficacy in psychotic patients, although there are differences in side effects. Hence, they are defined as "standard" as well as being typical with respect to parkinsonism. The results of an adequate trial of any one of them is considered representative of how the patient would be likely to respond to any other.3 Many psychiatrists used to believe that idiosyncratic responses may occur when switches are made among these six neuroleptics. Apart from better toleration due to differences in side effects, the literature does not offer any support for this impression. Perceived differences were often due to marked changes in dose (in chlorpromazine equivalents) that physicians made when switching from one to another.

To evaluate the adequacy of a trial of this type of medication, we rely on a clinical assessment of the central bioavailability of the drug. Pseudoparkinsonian side effects are considered to be the best readily available evidence of central bioavailability, as this is a marker for striatal dopamine type 2 (D2) receptor occupancy at a level of 75% to 80%. From numerous lines of evidence, this level of occupancy is above the minimally effective antipsychotic dose, and it is the best predictor of response to standard, typical neuroleptics.3,27 The absence of this clinical marker raises the question of whether the dose is adequate. However, the presence of this sign indicates an adequate dose and, furthermore, that there is little to be gained by raising the dose, and considerable risk of inducing more extrapyramidal effects.

Pseudoparkinsonism is diagnosed by the presence of the characteristic tremor described earlier or, more commonly, by the presence of cogwheel rigidity. Psychiatrists may use any examination process for detecting cogwheeling that they are accustomed to using. Perhaps the most sensitive examination procedure is the "neuroleptic threshold examination," which is discussed in detail in one of the "places to go" in the Internet program's desktop menu. This neurologic assessment requires only a minute or two of the physician's time. The examiner checks for cogwheel rigidity at the wrist joint during passive extension or flexion and circular movement, while the patient is distracted (to facilitate passivity) by skywriting with the other hand.26

Akathisia (restlessness or need to move limbs to relieve an uncomfortable sensation in the muscles) is not considered a sign or symptom of pseudoparkinsonism or of a bioavailable dose of an antipsychotic. Positron emission tomography scan studies suggest that akathisia occurs at approximately 60% to 65% occupancy of D2 receptors in susceptible patients. Thus, it often occurs below a dose that will be adequate for treating the psychosis.28

If there are no parkinsonian side effects, the current dose may not be bioavailable or adequate. Before this possibility is evaluated, it is worthwhile to review the medication treatment history. Has the patient had a trial of adequate duration with any standard, typical neuroleptic that produced parkinsonism and an unsatisfactory clinical response? If so, the patient may be considered a known nonresponder to this type of drug. The clinician can go ahead to the next section describing what to do next if the current neuroleptic trial is considered adequate.

Figure 3. Algorithm for completing an adequate trial of a standard, typical neuroleptic.

Figure 3. Algorithm for completing an adequate trial of a standard, typical neuroleptic.

To understand why there are no pseudoparkinsonian side effects from the current medication, the first question to ask is: Has the presenee of akathisia presented a barrier to raising the dose? If so, vigorous treatment of the akathisia might enable the dose to be raised. Figure 4 ("More #5") shows some steps to take in addressing akathisia, starting with an anticholinergic and going on to a beta-blocker, benzodiazepine, and various combinations. However, the development of akathisia, even if successfully treated, predicts a poor antipsychotic response to the neuroleptic.29 This patient will probably ultimately need to be switched to a different antipsychotic.

If akathisia is not a problem, we look for other reasons why there is no pseudoparkinsonism. If the patient is taking haloperidol, we first check a plasma level at this point, as that might establish compliance and, in addition, give some idea of whether there is enough drug in circulation. If the plasma level is below 5 ng / mL, which may be the threshold of a "therapeutic" level,30 we would raise it above 5 ng/mL (if akathisia does not prove to be a barrier) and see whether the patient responds. If it is currently above 5 ng/mL, we would conclude that there seems to be enough drug in circulation, but that it is not having the expected biologic effect in the brain (ie, parkinsonism). Figure 5 ("More #6") displays a strategy for this situation: first, address disabling akathisia, if any has developed, and then try pushing to higher blood levels of haloperidol up to a maximum of 20 ng/mL (if pseudoparkinsonian side effects do not appear first). This seems to be the upper limit of dosage at which, according to the literature, some treatment-resistant patients will finally respond.31 If the response remains unsatisfactory, we would declare this haloperidol trial to be adequate.

We would not use plasma levels to assist with the upward titration of the other five standard, typical neuroleptics. We might check a level to confirm compliance, however (Fig. 6, "Insufficient dose"). If compliance is not at issue, we would complete the trial by raising the dose to 1000 mg of chlorpromazine equivalents if tolerated, as was the standard in the Kane et al. clozapine study.32

It may seem somewhat heroic to push these neuroleptics to such high levels and doses. However, the lack of ability to generate parkinsonian side effects with standard, typical neuroleptics predicts poor subsequent response to clozapine.33 Thus, it could be worth the effort.

If there are parkinsonian side effects and the dose is therefore bioavailable, the clinician should return to the flowchart in Figure 3. The patient's trial is considered adequate if there has been a trial of 8 weeks (more or less, as described under treatment of first-onset outpatients). Again, there is considered to be no advantage to raising the dose over that which produces detectable pseudoparkinsonism. "Neuroleptic resistance" seems to have developed, in that clinically significant blockade of D2 receptors is not producing the expected benefits.34 There may be ways around the problem, such as through the use of atypical neuroleptics that may employ serotonergic mechanisms to regulate dopaminergic activity. At this point, however, the remaining intervention that might be worthwhile before declaring this neuroleptic trial to be adequate is to treat any akathisia that may be present. The akathisia may be producing disabling side effects and preventing an otherwise effective dose of neuroleptic from having its maximum effect.

If there is coarse pseudoparkinsonism, the neuroleptic dose might be lowered. This may be a common problem in Asian individuals and others with slow metabolism or pharmacodynamically based high susceptibility to neuroleptic side effects.35 The patient may be having an inhibited response from this excessive dose that is over the "window."

If the patient's symptoms rapidly worsen during the dose reduction, it might be a case of "supersensitivity psychosis." It may be helpful to add an anticonvulsant and try again to lower the dose.36 Alternatively, a switch to one of the newer, atypical antipsychotics is a reasonable option.363

This completes the algorithm for optimizing response to a standard neuroleptic. Figure 7 ("More #3") and Figure 8 show which antipsychotics to try next if the response has remained unsatisfactory. The logic "rules" of the algorithm may be summarized as follows: regardless of the sequence, we recommend two trials of new-generation antipsychotics and one trial of a standard, typical neuroleptic before consideration of clozapine. Our preference is for monotherapy in these trials.

Figure 4. Algorithm for akathisia.

Figure 4. Algorithm for akathisia.

2. Is the Patient Currently Taking One of the NewGeneration Antipsychotics- Olanzapine, Risperidone, or Quetiapine? If so, the first question is whether the patient has had an adequate trial of this type of medication Fig. 7). Parameters for adequate trials were presented earlier, in the section on first treatment with an antipsychotic. It may be reiterated that in the case of risperidone, clinical response is likely to be more favorable at doses that do not produce pseudoparkinsonism. However, if the patient does experience pseudoparkinsonism, risperidone is behaving like a standard, typical neuroleptic for that patient. The response pattern therefore may be more characteristic of a typical neuroleptic. If the response is unsatisfactory, this risperidone trial may be considered equivalent to a trial of a standard, typical neuroleptic ana may substitute for such a trial in the algorithm sequence.

The flowcharts in Figure 7 and Figure 8 now ask about a trial of another of the first-line, newgeneration antipsychotics, if this has not been done, and a trial of a standard, typical neuroleptic, if that has not yet been tried.

3. is the Patient Currently Taking One of the Older-Generation Antipsychotics That May Be Somewliat Atypical? In this category we consider the remaining available neuroleptics not induded in the "standard typical" list (ie, chlorprothixene, loxapine, mesoridazine besylate, molindone hydrochloride, pimozide, and thioridazine hydrochloride). The patient may be taking one of these at this time, and the question is whether the trial has been adequate. We recommend following the package insert (Food and Drug Adrninistration) recommendations for dosage, attempting to reach 1000 mg of chlorpromazine equivalents, if the dosage limits allow this and side effects are tolerable.

Figure 5. Algorithm for completing a haloperidol trial when there is no parkinsonism.

Figure 5. Algorithm for completing a haloperidol trial when there is no parkinsonism.

Chlorprothixene, molindone hydrochloride, and thioridazine hydrochloride are "somewhat atypical" in that they are relatively low in their ability to produce pseudoparkinsonism. Despite this, there is no support in the literature to suggest that they have any difference in clinical efficacy compared with the standard "typical" neuroleptics.3 Therefore, an adequate trial of one of these substitutes for a trial of a standard typical in the logic of the algorithm.

Loxapine, mesoridazine besylate, and pimozide are also "somewhat atypical" in the sense that they have biochemical or receptor activity profiles that make them unusual compared with the other neuroleptics. We will mention them again at another point. For the purpose of reaching rninimal eligibility for clozapine, the algorithm treats them as if they were standard, typical neuroleptics: if the trial has been adequate, this substitutes for a trial of a standard, typical neuroleptic.

Figure 6. Algorithm for insufficient dose of a standard neuroleptic. CPZ = chlorpromazine.

Figure 6. Algorithm for insufficient dose of a standard neuroleptic. CPZ = chlorpromazine.

4. Is the Patient Currently Taking Clozapine? If the patient is taking clozapine, our most effective antipsychotic in treatment-refractory cases, and the level of recovery is still unsatisfactory, it is first recommended that the dosage be optimized by adjusting the plasma level to be above 350 ng/mL, if side effects allow.37 The oral dose may be increased to 900 mg/d if necessary. This should be given in divided doses, at least during dosage adjustment, to lower the risks of seizures and respiratory compromise. When the level exceeds 350 ng/mL or the dose rises above 600 mg/d, it is recommended that an anticonvulsant such as valproate sodium be added to attempt prophylaxis of seizures.38

Here are some additional suggestions for enhancing the level of recovery in patients who are taking clozapine. The addition of low doses of typical neuroleptics or new-generation antipsychotics39,40 has some support, and this is often done when side effects prevent optimization of clozapine plasma levels.41 Obsessive-compulsive symptoms (which may be induced or exacerbated by clozapine and which contribute to the unsatisfactory response) may be addressed by a selective serotonin reuptake inhibitor42; however, the possible cytochrome P450 enzyme interactions with clozapine, especially with fluvoxamine43 (which is perhaps best avoided), must be considered. A mood component in schizoaffective disorder may respond to a mood stabilizer, but carbamazepine should be avoided due to its potential hematologic effects. Electroconvulsive therapy may be combined with clozapine with reasonable safety and with occasional marked benefit.4* Benzodiazepines may occasionally be added to clozapine for residual anxiety, but caution is advised due to reports of respiratory arrest, delirium, or orthostatic hypotension; most of these reports were associated with adding clozapine to an established benzodiazepine regimen, so therefore that sequence is not recommended.45

Figure 7. Algorithm for completing adequate trials of other antipsychotics.

Figure 7. Algorithm for completing adequate trials of other antipsychotics.

Step Three: Options to Consider If the Patient Is Eligible for Clozapine. If the response remains unsatisfactory after the patient has had two or three adequate monotherapy trials of appropriate antipsychotics as described above and all of the potential side effect issues have been addressed, clozapine is clearly an important option. Given clozapine's efficacy and effectiveness by a variety of measures, it should be employed sooner, rather than later, in the patient who is clearly treatment refractory.46 However, in some rural or understaffed community settings, there may be relatively inadequate support for optimal medical monitoring of clozapine. The psychiatrist or the patient may prefer to consider other options. These include the remaining antipsychotics, and adjunctive medications. Figure 8 ("More #4") displays these pathways.

Regarding the remaining antipsychotics, some algorithms recommend trying every new-generation antipsychotic before trying clozapine.47 That certainly may be reasonable for some patients, if notali.

In our earlier published reviews and algorithms, there was an educated guess that the "best of the rest" of the older generation of neuroleptics were the "somewhat atypicals" loxapine, mesoridazine besylate, and pimozide. This was based on evidence that may be considered weak, at best.3,48"52 This analysis was perhaps more interesting in the era prior to the availability of the new generation of antipsychotics, whose advantages are more clear. Nevertheless, we sometimes select these agents in preference to other standard, typical neuroleptics. In the case of pimozide, its usefulness seems limited to patients with treatment-resistant negative symptoms (eg, social withdrawal, blunted affect, paucity of speech, anhedonia, and apathy).49 Other strategies for negative symptoms are described below.

Figure 8. Options to consider if the patient meets criteria for clozapine. SSRI = selective serotonin reuptake inhibitor. TLE = temporal lobe epilepsy.

Figure 8. Options to consider if the patient meets criteria for clozapine. SSRI = selective serotonin reuptake inhibitor. TLE = temporal lobe epilepsy.

Although optional in this algorithm, and somewhat contrary to our general preference for minimizing polypharmacy, a trial of one or two adjunctive agents may be considered prior to clozapine for patients with an unsatisfactory response. Unfortunately, these adjunctives are often not particularly helpful,53 so the use of clozapine should not be unnecessarily delayed. The Consultant has much more detail on the rationale and current evidence supporting these various choices than has been included here. Choices are made based on which clinical subsyndrome is present that would comprise the target symptoms for the proposed adjunctive. These are briefly summarized below (Fig. 8).

Manic symptoms? Mood stabilizers: lithium, valproate sodium, carbamazepine, electroconvulsive therapy, or their combination if monotherapy provides partial response.

Impulsive, violent behavior? Lithium, carbamazepine, beta-blockers, or electroconvulsive therapy.

Depression, dysthymia? (occurring in the active phase of the psychosis) If the patient is depressed while taking a neuroleptic, switching to a new-generation antipsychotic is preferred.54,55 Antidepressants are of questionable benefit.56

Secondary ("post-psychotic") depression? Tricyclics and possibly other antidepressants.57

Residual negative symptoms? Be sure secondary negative symptoms are ruled out. Selective serotonin reuptake inhibitors,58 trazodone hydrochloride,59 d-cycloserine,60 glycine,61 or pimozide (refer to previous section).

Obsessive-compulsive symptoms? Clomipramine hydrochloride, possibly selective serotonin reuptake inhibitors.62

Other anxiety disorders? Selective serotonin reuptake inhibitors or nef azodone hydrochloride; benzodiazepines if no history of substance abuse.

Impairment of cognition? Research needed. Use newer antipsychotics. Ginkgo biloba extract?63

Specific organic factor contributing to the psychosis? Alzheimer's disease: carbamazepine, valproate sodium, or trazodone hydrochloride; history of severe hallucinogen abuse: carbamazepine; temporal lobe epilepsy: lithium or carbamazepine; hypercalcemic delirium: pimozide (a calcium channel blocker).64

THE CONSULTANT FOR THE PHARMACOTHERAPY OF SCHIZOPHRENIA: WHAT'S NEXT?

This computerized pharmacotherapy expert system has proposed strategies for the four clinical situations in which schizophrenic and schizoaffective patients most commonly present: first onset, relapse, agitation requiring parenteral treatment, and unsatisfactory level of response. In the course of a typical day for a busy physician, many questions arise regarding the optimal care of a patient that go unanswered.65 It could take hours to find answers in journals or texts, or by seeking an expert opinion. This software attempts to meet some of these information needs of the physician. However, this remains a work in progress. We welcome suggestions about how the algorithm could be improved or amplified to account for particular situations not addressed in the current version. Readers may give us this feedback electronically through the web site given at the begirining of the artide by clicking on "Send us a message."

REFERENCES

1. Osser DN. Treatment resistant problems. In: Tupin JP, Shader RI, Harnett DS, eds. Clinical Handbook of Psychopharmacology, 2nd ed. New York: Jason Aronson; 1988-269-328.

2. Osser DN. A systematic approach to pharmacotherapy in patients with neuroleptic-resistant psychoses. Hospital and Community Psychiatry. 1989;40:921-927.

3. Osser DN, Patterson RD. Pharmacotherapy of schizophrenia: 1. Acute treatment and 2. An algorithm for neuroleptic-resistant patients. In: Sorefi SM, ed. Handbook for the Treatment of the Seriously Mentally Rl. Toronto, Ontario, Canada: Hogrefe & Huber; 1996:91-155.

4. Zarate CA Jr, Cole JO. An algorithm for the pharmacological treatment of schizophrenia. Psychiatric Annals. 1994;24:333-340.

5. Zarate CA Jr. Algorithms for the diagnosis and management of schizophrenia and comorbid disorders. Primary Psychiatry. 1995;2:35-40.

6. Zarate CA Jr. Schizophrenia. In: Fawcett J, Stein DJ, Jobson KO, eds. Textbook of Treatment Algorithms in Clinical Psychopharmacology. Chichester, England: John Wiley and Sons; 1999:67-86.

7. Osser DN. The Neuroleptic-Resistant Psychosis Psychopharmacology Consultant: An Expert System. Lexington, MA: Mental Health Connections; 1991.

8. Shore MF. A lesson in history, circa 2047. Am J Psychiatry. 1997;154:307-311.

9. Conley RR, Kelly DL, Gale EA. Olanzapine response in treatment-refractory schizophrenic patients with a history of substance abuse. Schizophr Res. 1998;33:95-101.

10. Kirchner JE, Owen RR, Nordquist C, Fischer EP. Diagnosis and management of substance use disorders among inpatients with schizophrenia. Psychiatr Serv. 1998;49:82-85.

11. Sanger T, Lieberman JA, Tohen M, Grundy S, Beasley C Jr, Tollefson GD. Olanzapine versus haloperidol treatment in first-episode psychosis. Am J Psychiatry. 1999;156:79-87.

12. Kopala LC, Good K, Honer WG. Extrapyramidal signs and clinical symptoms in first episode schizophrenia: response to low-dose risperidone. / Clin Psychopharmacol. 1997;17:308-313.

13. Thomas CS. Which atypical antipsychotic? Br J Psychiatry. 1998;172:106-109.

13a. Robinson DG, Woerner MG, Alvir JMI, et al. Predictors of treatment response from a first episode of schizophrenia or schizoaffective disorder. Am ) Psychiatry. 1999;156:544549.

14. Beasley CMJ, Tollefson G, Tran P, Satterlee W, Sanger X Hamilton S. Olanzapine versus placebo and haloperidol: acute phase results of the North American double-blind olanzapine trial. Neuropsychopharmacology. 1996;14: 111-123.

15. Osser DN, Najarian D, Dufresne RL. Olanzapine increases weight and triglycerides. J Clin Psychiatry. 1999. In press.

15a. Dickson RA. Antipsyehotic-induced hyperprolactinemia causes serious adverse effects: pro. 77ie journal of Psychotic Disorders. 1997;2:3-15.

16. Bassett A, Addington D. Canadian clinical practice guidelines for the treatment of schizophrenia. Can J Psychiatry. 1998;43:29S.

17. Brecher M, Jeste D, Okamoto A, Napolitano J, Kane JM. Low frequency of tardive dyskinesia in elderly patients with dementia. American College of Neuropsychopharmacology 37th Annual Meeting; Las Croabas, Puerto Rico; 1998.

18. Steingard S, Allen M, Schooler NR. A study of the pharmacologic treatment of medication compliant schizophrenics who relapse. J Clin Psychiatry. 1994;55:470472.

19. Fenton WS, BIyler CR, Heinssen RK. Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr Bull. 1997;23:637-651.

20. Love RC, Conley RR, Kelly DL, Bartko JJ. A comparison of rehospitalization rates between patients treated with atypical antipsychotics and those treated with depot antipsychotics (Abstract 1-32). American College of Neuropsychopharmacology 37th Annual Meeting; Las Croabas, Puerto Rico; 1998.

21. Sperry L. Psychopharmacology and psychotherapy: strategies for maximizing treatment outcomes. In: Sauber SR, ed. Mental Health Practice Under Managed Care. New York: Brunner/Mazel; 1995:193.

22. Hymen S. Acute psychoses and catatonia. In: Hyman S, Tesar G, eds. Manual of Psychiatric Emergencies, 3rd ed. Boston: Little, Brown; 1994.

23. Thomas HJ. Droperido] versus haloperidol for chemical restraint of agitated and combative patients. Ann Emerg Med. 1992;21:407-413.

24. Tandon R, DeQuardo JR, Goodson J, Mann NA, Greden JR Effect of anticholinergics on positive and negative symptoms in schizophrenia. Psychopharmacol Bull. 1992;28:297-302.

25. van Putten T, Marder S, Mintz J, Poland R. Plasma fluphenazine levels and clinical response in newly admitted schizophrenic patients. Psychopharmacol Bull. 1991;27:91-96.

26. Osser DN. Neuroleptic-induced pseudoparkinsonism. In: Joseph AB, Young RR, eds. Movement Disorders in Neurology and Neuropsychiatry, 2nd ed. Boston: Blackwell; 1999:61-68.

27. Kapur S, Zipursky RB, Remington G. Clinical and theoretical implications of 5-HT2 and D2 receptor occupancy of clozapine, risperidone, and olanzapine in schizophrenia. Am ] Psychiatry. 1999;156:286-293.

28. Nordstrom A-L, Farde L, Wiesel F-A. Central D2dopamine receptor occupancy in relation to antipsychotic drug effects: a double-blind PET study of schizophrenic patients. Biol Psychiatry. 1993;33:227-235.

29. Levinson DF, Simpson GM, Singh H, et al. Fluphenazine dose, clinical response, and extrapyramidal symptoms during acute treatment. Arch Gen Psychiatry. 1990;47: 761-768.

30. van Putten T, Marder S, Mintz J, Poland R. Haloperidol plasma levels and clinical response: a therapeutic window relationship. Am J Psychiatry. 1992;149:500-505.

31. Coryell W, Miller DD, Perry PJ. Haloperidol plasma levels and dose optimization. Am } Psychiatry. 1998;155:48-53.

32. Kane JM, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry. 1988;45:789-796.

33. Lieberman J, Safferman AZ, Pollack S, et al. Clinical effects of clozapine in chronic schizophrenia: response to treatment and predictors of outcome. Am ] Psychiatry. 1994;151:1744-1752.

34. Meltzer HY, Lee M; Cola P. The evolution of treatment resistance: biologic implications. J Clin Psychopharmacol. 1998;18(suppl 1):5S-11S.

35. Pi EH, Gray GE. A cross-cultural perspective on psychopharmacology. Directions in Psychiatry. 1998,18: 425-441.

36. Chouinard G, Sultan S. Treatment of supersensitivity psychosis with antiepileptic drugs: a report of a series of 43 cases. Psychopharmacol Bull. 1990;26:337-341.

36a. Chouinard G, Beaudry P, Vainer J, Turnier L, Roy J-Y, Miller R. Risperidone and clozapine in the treatment of neuroleptic-induced supersensitivity psychosis. Biol Psychiatry. 1993;33:130A.

37. VanderZwaag C, McGee M, McEvoy JP, Freudenreich O, Wilson WH, Cooper TB. Response of patients with treatment-refractory schizophrenia to clozapine within three serum level ranges. Am ) Psychiatry. 1996;153:1579-1584.

38. Freudenreich O, Weiner RD, McEvoy JP. Clozapineinduced electroencephalogram changes as a function of dozapine serum levels. Biol Psychiatry. 1997;42:132-137.

39. Gupta S, Sonnenberg SJ, Frank B. Olanzapine augmentation of clozapine. Ann Clin Psychiatry. 1998;10:113-115.

40. Henderson DC, Goff DC. Risperidone as an adjunct to clozapine therapy in chronic schizophrenics. J Clin Psychiatry. 1996;57:395-397.

41. Peacock L, Gerlach J. Clozapine treatment in Denmark: concomitant psychotropic medication and hematologic monitoring in a system with liberal usage practices. } Clin Psychiatry. 1994;55:44-49.

42. Allen L, Tejera C. Treatment of clozapine-induced obsessive-compulsive symptoms with serbaline. Am / Psychiatry. 1991;151:1096-1097. Letter.

43. Shader Rl, Greenblatt DJ. Clozapine and fluvoxamine: a curious complexity. / Clin Psychopharmacol. 1998;18: 101-102.

44. Klapheke MM. Combining ECT and antipsychotic agents: benefits and risks. Convulsive Therapy. 1993;9:241-255.

45. Faisal I, Lindenmayer JP Taintor Z, Cancro R. Clozapinebenzodiazepine interactions. / Clin Psychiatry. 1997;58: 547-548. Letter.

46. Bradford DW, Chakos MH, Sheitman BB, Lieberman JA. Atypical antipsychotic drugs in treatment- refractory schizophrenia. Psychiatric Annals. 1998;28:618-626.

47. Chiles JA, Miller AL, Crismon ML, Rush AJ, Krasnoff AS, Shon SS. The Texas Medication Algorithm Project: development and implementation of the1 schizophrenia algorithm. Psychiatr Serv. 1999;50:69-74.

48. Tuason VB, Escobar JI, Garvey M, Schiele B. Loxapine versus cWorpromazine in paranoid schizophrenia: a doubleblind study. J Clin Psychiatry. 1984;45:158-163.

49. Opler LA, Feinberg SS. The role of pimozide in clinical psychiatry: a review. / Clin Psychiatry. 1991;52:221-233.

50. Osser DN, Albert LG, Figueiredo S, O'Connor H, Barden Y, Carmichael WG. Mesoridazine in neuroleptjc-resistant psychoses. / Clin Psychopharmacol. 1991;11:328-330.

51. Kapur S, Zipursky R, Remington G, Jones C, McKay G, Houle S. PET evidence that loxapine is an equipotent blocker of 5-HT2 and D2 receptors: implications for the therapeutics of schizophrenia. Am ] Psychiatry. 1997;154:1525-1529.

52. Bishop MP, Simpson AM, Dunnett CW. Efficacy of loxapine in the treatment of paranoid schizophrenia. Psychopharmacology. 1977;51:107-115.

53. Christinson GW, Kirch DG, Wyatt RJ. When symptoms persist: choosing among alternative somatic treatments for schizophrenia. Schizophr Bull. 1991;17:217-245.

54. Tollefson GD, Sanger TM, Thieme ME. Depressive signs and symptoms in schizophrenia: a prospective blinded trial of olanzapine and haloperidol. Arch Gen Psychiatry. 1998;55:250-258.

55. Cantillon M, Goldstein JM. Efficacy of quetiapine in affective symptoms of schizophrenia (NR 445). American Psychiatric Association Annual Meeting; Toronto, Ontario, Canada; 1998.

56. Lee MS, Yong KK, Lee SK, Suh KY. A double-blind study of adjunctive sertraline in haloperidol-stabilized patients with chronic schizophrenia / Clin Psychopharmacol. 1998;18:399-403.

57. Plasky P. Antidepressant usage in schizophrenia. Schizophr Bull. 1991;17:649-656.

58. Goff DC, Midha KK, Sarid-Segal O, et al. A placebo-controlled trial of fluoxetine added to neuroleptic in patients with schizophrenia. Psychopharmacology. 1995;117:417-423.

59. Decini P, Mukherjee S, Bocola V, Saraceni F, Hadjichristos C, Scapicchio P. Adjunctive trazodone in the treatment of negative symptoms of schizophrenia. Hospital and Community Psychiatry. 1994;45:1220-1223.

60. Goff DC Tsai G, Levitt J, et al. A placebo-controlled trial of d-cycloserine added to conventional neuroleptics in patients with schizophrenia. Arch Gen Psychiatry. 1999;56:21-27.

61. Heresco-Levy U, Javitt DC, Ermilov M, Mordel C, Silipo G, Lichtenstein M. Efficacy of high dose glycine in the treatment of enduring negative symptoms of schizophrenia. Arch Gen Psychiatry. 1999;56:29-36.

62. Sasson Y, Bermanzohn PC, Zohar J. Treatment of obsessive-compulsive syndromes in schizophrenia. CNS Spectrums. 1997;2:34-36, 45.

63. Berman L Pappas D, Levinthal N, Sigadel R, Patel CK. Ginkgo biloba extract in schizophrenic patients (NR 259). American Psychiatric Association Annual Meeting; San Diego, California; 1997.

64. Mark B, Shakin-Kunkel E, Fabi M, Thompson Tl. Pimozide is effective in delirium secondary to hypercalcemia when other neuroleptics fail. Psychosomatics. 1993;34:446-450.

65. Westberg EE, Miller RA. The basis for using the Internet to support the information needs of primary care. } Am Med Inform Assoc. 1999;6:6-25.

10.3928/0048-5713-19990501-08

Sign up to receive

Journal E-contents