Additional related information may be found at:
Neuropsychopharmacology: The Fifth Generation of Progress

Back to Psychopharmacology - The Fourth Generation of Progress

Atypical Antipsychotic Drugs

Herbert Y. Meltzer


The distinction between typical and atypical antipsychotic drugs has attained considerable importance since the last edition of this book, because of evidence, reviewed below, that a new generation of antipsychotic drugs with important advantages over the first generation of antipsychotic drugs (i.e., the "typical" antipsychotics, usually referred to as neuroleptics) has been introduced into clinical practice or is in an advanced stage of development (44). Ironically, the best-studied atypical antipsychotic drug is not a new drug at all but an old one, clozapine. Although first synthesized in 1959, the scope and importance of the differences between clozapine and atypical neuroleptics was not appreciated until 1988, following the demonstration of its greater efficacy in at least 30% of schizophrenic patients who had failed to respond to at least three trials of neuroleptic drugs of different classes (27). Other atypical antipsychotic drugs have not been as well studied as clozapine, but there is enough evidence of their relationship to clozapine and of clinically important differences between them and neuroleptics to warrant the new designation.

It should be noted that not all investigators approve of the designation of a group of atypical antipsychotic drugs, preferring to describe all antipsychotic agents in terms of their currently perceived key pharmacological properties (22). This is, in part, because of a lack of agreement on what is the essential definition of atypicality in an antipsychotic drug (42). Common to all definitions of atypical antipsychotic drugs of which I am aware is the ability to produce an antipsychotic action in most patients at doses that do not cause significant acute or subacute extrapyramidal side effects (EPS), such as parkinsonism and akathisia. By this definition, remoxipride, a substituted benzamide, is atypical (32), as is risperidone, a benzisoxazole, although the latter drug's low EPS profile is evident only at lower doses (e.g., 6–8 mg/day) (9, 14). At higher doses, risperidone may be indistinguishable from haloperidol with regard to EPS, indicating that the distinction between typical and atypical antipsychotics requires careful study of dose–response relationships for both EPS and efficacy. Other characteristics that have been included in some definitions of atypicality include (a) failure to increase serum prolactin (PRL) levels; (b) superior efficacy for positive, negative, and disorganization symptoms; and (c) no evidence of tardive dyskinesia or dystonia following chronic administration. This second cluster of properties has been shown to be characteristic of clozapine (43), which is why they are considered by some to be criteria for atypicality. There is evidence that some antipsychotic drugs that satisfy the low acute EPS criterion meet some, but not all of these criteria (e.g., both risperidone and remoxipride elevate serum PRL levels) (72, 78). The lower rate and intensity of acute and subacute EPS produced by different atypical antipsychotics may be the result of a variety of biological properties, which they may not share. The differences in clinical profile among atypical antipsychotics identified solely on the basis of low EPS suggests that multiple biological mechanisms may be responsible for the diverse differences between clozapine and typical antipsychotic drugs.

This chapter reviews the recent studies of efficacy, side effects, and indications for clozapine, risperidone, and remoxipride, the other atypical antipsychotic drugs that have been approved for use in a number of countries. Other strategies for developing atypical antipsychotic drugs are also briefly discussed (see also Neurophysiological and Psychophysiological Approaches to Schizophrenia and Neurocognitive Functioning in Paitents with Schizophrenia: An Overview ).


Clozapine, a dibenzazepine tricyclic that is chemically related to loxapine, which is a neuroleptic, was first clinically tested in the 1960s in Europe. The preclinical profile of clozapine is consistent with a drug that would be an effective antipsychotic with few EPS (i.e., it does not produce catalepsy but does block amphetamine-induced locomotor activity) (42). Its effectiveness as an antipsychotic without producing EPS was confirmed in early clinical trials (3). However, it was observed to produce agranulocytosis at a rate much higher than that usually found with standard neuroleptic drugs (29), approximately 1 in 2,000. This led to the withdrawal of clozapine from clinical use in Europe and the cessation of widespread testing in the United States. However, when many of the schizophrenic patients withdrawn from clozapine at that time did not respond as well to typical neuroleptic drugs or expressed strong preference for clozapine because of its low EPS profile, they were permitted to continue receiving clozapine for prolonged periods. This led to a body of clinical information unparalleled in the history of psychotropic drugs which lack formal approval for general use in a developed society. Clinical experience with clozapine suggested that it differed from typical neuroleptic drugs in causing fewer EPS, had a greatly diminished liability to cause tardive dyskinesia, and that it possibly had superior efficacy for some patients with schizophrenia. Definitive evidence that it did not cause elevations of serum PRL concentrations in man was also noted (42).

A double-blind six-week trial comparing clozapine with chlorpromazine in hospitalized neuroleptic-resistant patients established that is was superior to chlorpromazine for alleviating both positive and negative symptoms as well as improving ward behavior (27). This study also showed that clozapine was well-tolerated in a small group of neuroleptic-intolerant patients. On the basis of this study, clozapine was approved for use in the United States and subsequently in other countries.

Because of its ability to cause granulocytopenia or agranulocytosis in 1% to 2% of patients (29), clozapine has usually not been considered to be a first line drug for the initial treatment of schizophrenia or for those chronic patients whose positive symptoms respond to typical antipsychotic drugs (see below). Rather, the chief indication for clozapine is considered to be neuroleptic-resistant schizophrenic patients (i.e., those patients with schizophrenia who have an unsatisfactory response to at least two trials of typical neuroleptic drugs of adequate dosage and duration) (27, 43). There appears to be no need for three trials if two are unsuccessful. These patients are characterized by persistent moderate positive, negative, or disorganization symptoms, along with impaired cognitive and social function. Conservatively, at least 30% of schizophrenic patients are neuroleptic-resistant. Some schizophrenic patients are resistant to neuroleptics from the first episode, whereas other patients become resistant at a later phase of their illness (38). Pharmacokinetic factors do not account for the poor response to neuroleptic drugs as positron emission studies using D2 ligands such as 11C-N-methylspiperone demonstrate adequate occupancy of central D2 receptors in neuroleptic-resistant patients.

The minimum standard for an unsatisfactory response to neuroleptic drugs has not been established. If even moderate social impairment, persistent negative symptoms, and mild-to-moderate positive symptoms are present despite treatment with neuroleptics, it would be reasonable to consider clozapine, because clozapine should produce clinically significant benefits in the majority of such patients.

Patients that are neuroleptic intolerant (i.e., patients with tardive dyskinesia of at least moderate severity despite optimal adjustment of neuroleptic dosage) and patients who cannot tolerate therapeutic doses of even low EPS-producing antipsychotic drugs, such as thioridazine, and are frequently noncompliant because of it, should also be considered candidates for clozapine (43).

The criteria for use of clozapine will change when other drugs are approved that possess some or all of clozapine's advantages. Trials of these agents may be expected to precede trials with clozapine if the risk–benefit ratio of these newer drugs for specific indications is superior. This may be the case for risperidone in neuroleptic-intolerant patients. The risk–benefit ratio for clozapine treatment of neuroleptic-responsive schizophrenic patients must be developed in controlled trials that evaluate multiple dimensions of outcome, including psychopathology, cognition, social function, work function, hospitalization, and cost effectiveness.

There is limited clinical evidence that clozapine is effective in neuroleptic-resistant psychotic children and young adolescents, as well as in elderly, neuroleptic-resistant, or intolerant schizophrenic patients (43). Clozapine is also useful in alleviating psychotic symptoms in demented psychogeriatric patients.

Twelve studies have compared the efficacy of clozapine with standard neuroleptic drugs in non-treatment-resistant schizophrenia (see ref. 3 for a review of these studies). Clozapine was reported to be superior to typical neuroleptic drugs with regard to global psychopathology or specific positive symptoms in seven of these studies whereas five studies failed to show a difference between clozapine and typical drugs. Differences in dosage, duration of study, sample size, and patient composition may account for these differences.

Other Indications

There is limited evidence from case reports that clozapine may be at least as, or more, effective than other antipsychotics in a variety of conditions, such as treatment-resistant mood disorders, including: (a) rapid cycling and dysphoric mania (13); (b) psychotic depression (59); (c) organic psychoses, such as cases of Huntington's chorea refractory to neuroleptic drugs; and (d) refractory psychoses due to severe head injury (51). Clozapine has also proven useful to treat polydypsia in schizophrenic patients (31). Clozapine, at low doses, is extremely useful in treating L-dopa-induced psychoses (19). It seems reasonable that clozapine should be evaluated for use in refractory patients with a variety of neuropsychiatric disorders in which neuroleptic drugs have been shown to be therapeutic.


Initiation of clozapine treatment should ordinarily be done in patients free of other psychotropic drugs to minimize side effects such as hypotension, sedation, and anticholinergic effects and to avoid interference with those benefits of clozapine that are dependent upon weak D2-receptor blockade (e.g., low EPS and possibly increased efficacy and cognitive improvement) (43). If typical neuroleptic drugs are coadministered with clozapine to initiate treatment, they should be stopped within the first 2 to 3 weeks. However, controlled trials of the combination of clozapine and neuroleptic drugs versus clozapine alone, are needed to establish whether the use of a neuroleptic does, in fact, delay or impair response to clozapine.

The recommended starting dose of clozapine is 12.5 mg to test for possible hypotensive reactions. Slow titration is recommended until doses of 300 to 450 mg/day are reached, generally by 2 to 3 weeks. This may be done on an outpatient basis. Optimal doses usually range between 450 and 600 mg/day in younger adults, but doses up to 900 mg/day may be needed. Twice-a-day dosage is recommended because the half-life of clozapine is 12 to 16 hr (27). Elderly patients usually respond to 200 to 300 mg/day or even lower doses. There is no data available as to whether lower doses of clozapine may be effective for maintenance treatment.

There are, as yet, no fixed-dose studies to determine optimal dosage. In Europe, clinical practice has been to use doses of 200 to 300 mg/day or even lower (55), whereas in the United States and England, doses of 400 to 600 mg/day are most common (43). The reasons for this discrepancy require further study. It is possible that lower doses are required for neuroleptic-responsive or neuroleptic-intolerant patients than for neuroleptic-resistant cases.

Clozapine improves the three main syndromes of psychopathology in schizophrenia: positive symptoms, negative symptoms, and disorganization (27, 45). Improvement is greatest in disorganization followed by positive and then negative symptoms. Although the change in negative symptoms may be slight (10), it can be shown to be independent of the change in positive symptoms, EPS, and depression (52). Affective symptoms in schizoaffective patients improve concurrently. Clozapine appears to be effective in treating aggression and hostility in schizophrenic patients (76). Improvement in social functioning during clozapine treatment has been reported (46). Even some previously regressed patients with marked defect symptoms have been able to return to work after clozapine (36, 46). The marked reduction in rehospitalization that usually results from clozapine treatment also contributes to improvement in social function (42). Psychosocial treatments have been advocated as necessary to facilitate recovery during clozapine treatment (43).

One of the major advantages of clozapine is the ability to improve some aspects of cognitive dysfunction in schizophrenia (23) Cognitive dysfunction is characteristic of schizophrenia from the first episode and may not differ between neuroleptic-resistant and intolerant cases. Typical antipsychotics have rarely been found to produce significant improvement in any aspect of cognitive function. Clozapine has been found to have modest effects to improve attention, verbal fluency, recall memory, and executive function (23) It does not improve performance on the Wisconsin Card Sort Test (23). It remains to be determined if other atypical antipsychotic drugs can also improve cognitive function.

Response to clozapine in neuroleptic-resistant patients may not be evident until after 6 months or after even longer treatment periods (41). Approximately 30% respond by 6 weeks and another 30% respond more slowly. The reasons for such a delayed response to clozapine are unknown. Because of the frequent delay in the onset of its effect and its potential for clinical superiority to standard drugs, clozapine should probably not be discontinued for lack of apparent efficacy before less than six months of treatment have elapsed.

Clozapine may be no more effective than standard neuroleptics in treating positive symptoms in up to 25% of neuroleptic-resistant schizophrenic patients, despite a trial of adequate dose (up to 900 mg/day) or duration (up to 6 months or longer) (46, 47, 48). Addition of typical neuroleptic drugs or electroconvulsive therapy (ECT) has been reported to be helpful in some cases, but controlled data is lacking. Selective serotonin uptake inhibitors such as fluoxetine and paroxetine have proven useful to alleviate persistent or emergent depression or obsessive–compulsive symptomatology that is sometimes increased by clozapine (2).

Abrupt withdrawal of clozapine may sometimes be associated with a rapid, severe, and sometimes prolonged exacerbation of psychotic symptoms (58), but this is relatively rare. When clozapine treatment is terminated because of noncompliance or for some other reason and then reinstituted, the response is usually similar but some cases of greatly diminished responsivity have been reported. The reason for this is unknown.

Clinical Pharmacology and Plasma Levels

Peak clozapine plasma levels following the oral administration of clozapine are achieved in 1 to 4 hr (26). The half-life after twice-a-day dosing at the steady state is approximately 14 hr (range 6 to 33 hr). Thus, steady-state levels should be achieved approximately 1 week after constant dose twice-daily administration. The major metabolites of clozapine are desmethylclozapine and clozapine-N-oxide; but they appear to be inactive.

Optimal plasma levels of clozapine in treatment-resistant schizophrenia have been reported to be ³ 350 ng/ml, range 60 to 1000 ng/ml (24, 60), but some patients respond at lower levels. There is no evidence for a therapeutic window. It may be prudent in very carefully selected cases to exceed the recommended limit of 900 mg/day if plasma levels of clozapine are less than 350 ng/ml, and there are no major cardiovascular or other side effects (24). Clozapine metabolism may be blocked by cimetidine, phenytoin, valproic acid, and fluoxetine (see 43 for a review).

Prefrontal sulcal prominence as measured by CT has been reported to be inversely related to response to clozapine, but ventricular brain ratio (VBR) does not predict clinical response (19). Shorter duration of illness, female gender, and younger age are significant predictors of a more favorable response to clozapine.

Drug Interactions

There has been considerable concern about combining clozapine and benzodiazepines because of possible respiratory arrest. However, reliable evidence for a negative interaction between clozapine and benzodiazepines is very slight. Benzodiazepines are sometimes useful in diminishing anxiety when initiating clozapine treatment in neuroleptic-free schizophrenic patients (34).

The combination of clozapine and lithium has been implicated in several cases of neuroleptic malignant syndrome (61) as well as in other types of neurotoxicity (8). These reports suggest lithium should be used with clozapine only if hypomanic symptoms are not adequately controlled with clozapine. Valproic acid may be the safest agent to combine with lithium carbonate therapy if supplemental mood stabilization is required. However, valproic acid may also increase clozapine plasma levels.

Side Effects

Granulocytopenia and Agranulocytosis

The major side effect of clozapine, which limits its use, is its significant risk of granulocytopenia or agranulocytosis (1, 29). In the United States, indefinite weekly total white blood cell (WBC) or neutrophil counts are required to permit rapid detection of this potentially fatal side effect. Some countries require only monthly monitoring after the first 4 to 6 months of treatment, because approximately 50% to 80% of cases of neutropenia or agranulocytosis occur within the first 18 weeks of treatment.

The fall in WBC count due to clozapine may be abrupt or gradual. Steadily falling total WBC or neutrophil counts should increase concern, even if none are below 3000/mm3. An abrupt drop of 3000/mm3 or more in a single week may signal impending agranulocytosis. When the total WBC count falls below 3000/mm3 or the neutrophil count below 1500/mm3, clozapine must be discontinued and the white count with differential must be followed for 4 weeks. Patients who are withdrawn from clozapine during a neutropenic phase (WBC count between 2500 and 3500/mm3), may be rechallenged, but more intensive and differential monitoring is indicated.

The etiology of clozapine-induced agranulocytosis is unknown. Clozapine itself or its metabolite desmethylclozapine have been implicated in the etiology of agranulocytosis or granulocytopenia (21). Because the duration of agranulocytosis may be longer than for other drug-induced blood dyscrasias, neutropenia fever, should it develop, is often a challenging management problem. Upon withdrawal of clozapine, the hematological status usually returns to normal within 2 to 3 weeks. The successful use of granulocyte colony stimulating factor (G-CSF) to promote more rapid recovery from agranulocytosis has recently been reported (4).

Patients who have developed agranulocytosis on clozapine will experience it again, after only 1 to 4 weeks of treatment (29). Thus, rechallenge with clozapine after clear granulocytopenia or agranulocytosis should not be tried.

Extrapyramidal Reactions

The absence, or very low incidence, of EPS is a major clinical advantage associated with clozapine therapy (43) and contributes significantly to patient acceptance and compliance. There have been no reported cases of dystonic reactions in patients receiving clozapine monotherapy. Compared to typical neuroleptic drugs, clozapine produces much less akathisia. It does not worsen parkinsonian symptoms in Parkinson's disease and may even produce some improvement in bradykinesia, tremor, and rigidity (19). There have been no definite reported cases of tardive dyskinesia linked to clozapine treatment alone. At the same time, clozapine appears effective in blocking this syndrome in up to 67% of patients (12, 35). Here again response may be delayed. Approximately 30% of cases with tardive dyskinesia (TD) treated with clozapine will have a remission of TD and another 30% will have a reduction in severity. However, symptoms recur when clozapine is stopped. Clozapine can cause neuroleptic malignant syndrome, which should be managed in the same manner as when the etiology is a typical neuroleptic.

Other Side Effects

Clozapine decreases the seizure threshold, sometimes causing a dose-related increase in major motor seizures or myoclonus (18). Major motor seizures are induced in 1% to 2% of patients at doses below 300 mg/day, 2% to 4% at doses greater than 300 mg/day, and 4% to 6% risk at doses greater than 600 mg/day. Management of these seizures is usually possible, so that discontinuation because of seizures is rarely necessary. Decreasing the dose and pharmacological management of seizures (e.g., addition of valproic acid) is usually able to control the seizures. The electroencephalogram (EEG) is usually abnormal in clozapine-treated patients and is, therefore, of little clinical value (37).

There have been numerous reports of clozapine-induced exacerbations of obsessive–compulsive symptoms (2). This may be due to serotonin2 (5-HT2) receptor blockade and usually responds to treatment with selective serotonin reuptake inhibitors (SSRIs).

The main cardiovascular side effects of clozapine are orthostatic hypotension, tachycardia, and some conduction abnormalities (34). Tachycardia, in patients receiving clozapine, is predominantly the result of the anticholinergic effects of clozapine. Mean heart rate can increase 20 to 25 beats per minute and can persist for over a year in some patients. Similar to orthostatic hypotension, this effect is also dose dependent. Tolerance to the tachycardia develops slowly. b-Adrenergic blockers have been successfully used to decrease clozapine-induced tachycardia. Hypertension may also be observed, most frequently in the first 2 weeks of treatment.

Hypersalivation occurs in about 30% of clozapine-treated patients. Reduction in the dosage of clozapine or treatment with an anticholinergic medication such as benztropine can prove beneficial (34). The a2-adrenergic agonist clonidine, which causes dry mouth by suppressing sympathomimetic sialagogic mechanisms, may also be helpful in treating hypersalivation due to clozapine.

Weight gain is common with clozapine. In a recent study, the mean clozapine-induced weight gain was 6 kg or 8.9% of body weight (30). The magnitude of weight gain has been reported to be positively correlated with clinical response.

Clozapine does not elevate serum PRL levels (43).

Cost Effectiveness

Two studies have examined the cost effectiveness of clozapine in treatment-resistant schizophrenia. Revicki et al. (63) reported that by the second year of treatment, total costs had decreased more than $24,000 per year in 86 treatment-resistant schizophrenic patients. In a prospective study, Meltzer et al. (47) compared the total costs of treatment for 2 years before and 2 years after clozapine treatment in 37 treatment-resistant schizophrenic patients. The median total cost for 2 years decreased from $42,934 (1987 dollars) to $23,772, a decrease of $19,162. These studies suggest that large savings are possible with clozapine treatment for treatment-resistant schizophrenic patients who have had frequent hospitalizations.


Clozapine provides the possibility of significant help for at least 60% of patients with schizophrenia who fail to respond adequately to typical neuroleptics, as well as a reduced potential to cause EPS, including TD. On the other hand, because of its side-effect profile and the need for blood monitoring, the use of clozapine has been relatively limited. As of April, 1994, it has been tried on 65,000 of the approximately 2,000,000 schizophrenic patients in the United States. Its fate will depend upon the ability of other novel antipsychotic agents, described below, to have similar advantages in efficacy, EPS, and TD, without other serious adverse effects.


Risperidone is a benzisoxazole derivative, chemically unrelated to any other currently available antipsychotic drug. Risperidone, like clozapine, is a potent 5-HT2A, 5-HT7, a1-, a2-adrenergic and histamine H1 and a relatively weak (in comparison with its affinity for the 5-HT2 receptor) D2 dopamine (DA) receptor antagonist (34). However, its absolute affinity for the D2 receptor is similar to that of haloperidol (1 to 5 nM). It has low potency as an antagonist at D1 and D4 receptors. In vivo, risperidone blocks DA agonist-induced locomotor activity at doses that cause weak catalepsy, a profile that suggests it should produce less EPS at effective antipsychotic doses than typical neuroleptics (40). Risperidone, unlike clozapine, stimulates prolactin secretion in man (6, 73), suggesting that at clinically effective doses, it produces a net decrease in dopaminergic activity at the pituitary lactotrophs and should cause galactorrhea in some females.

Clinical Pharmacology

Risperidone is rapidly and completed absorbed following oral administration and reaches peak plasma levels within 2 hr (73). It is extensively metabolized in the liver; 9-hydroxyrisperidone, the major metabolite, has a half-life of 17 to 22 hr compared to 2 to 4 hr (mean 2 to 8 hr) for risperidone. Thus, the pharmacologically active moiety is the sum of risperidone and 9-hydroxyrisperidone. Poor metabolizers of debrisoquine (8% of the Caucasian population) will also be unable to metabolize risperidone because the 9-hydroxylation of risperidone is catalyzed by CYP2 D6. Poor metabolizers of risperidone have half the active moiety of good metabolizers because risperidone levels will be much higher in the poor metabolizers. Elimination of the active moiety is independent of metabolic phenotype with a half-life of 20 hr. For this reason, poor metabolizers of risperidone require the same daily dosage as good metabolizers. There is no data about a relationship between plasma levels of the active moiety and clinical response. No parenteral formulation, either short- or long-acting, is currently available.


Risperidone, in doses of 4 to 10 mg/day, appears to be at least equivalent and possible superior to haloperidol, 10 to 20 mg/day, in decreasing positive and negative symptoms (6, 9, 13, 14, 16, 50). For some patients, slightly lower or higher doses may be needed. Some studies have reported risperidone to be superior to haloperidol in treating positive and negative symptoms in schizophrenia and to possibly have a faster onset of action (14, 16). Risperidone was found to be superior to placebo as add-on therapy to reduce behavioral disturbances in mentally retarded patients (72). There is no evidence yet as to its effect on cognition. The long-term benefits of risperidone and its cost-effectiveness has not yet been studied. The proportion of patients for whom risperidone have superior efficacy to standard treatment is not entirely clear and needs further study. Two of the above studies (14, 16) included some apparently neuroleptic-resistant schizophrenic patients, raising the possibility that risperidone might be effective in some, if not all, the patients for whom clozapine is effective. However, no double-blind comparison of risperidone and clozapine has been conducted to date, and, until definitive evidence is available, it would be premature to broaden the indication for risperidone to this class of patients, except for experimental purposes or for patients who fail or cannot tolerate clozapine.

There is some evidence that risperidone can partially mask TD symptoms (14). It has been suggested that risperidone has a lower potential to cause TD or dystonia than typical neuroleptic drugs because it produces fewer EPS, but this is not necessarily the case, as indicated by thioridazine.

Side Effects

Risperidone has been reliably reported to produce fewer EPS than haloperidol at doses of 4 to 6 mg/day (9). Consistent with this, antiparkinsonian agents are needed less frequently with risperidone than with haloperidol but may be required by approximately 20% of patients (14, 16). There is clear evidence that the incidence of EPS with risperidone is dose-related, and it should be comparable to haloperidol at higher doses ({ewc MVIMG, MVIMAGE,!greateq.bmp}12 mg/day). In clinical practice, it is important to keep the dose of risperidone in the 4 to 8 mg/day range wherever possible. This may require lowering the dose after the acute phase of psychosis is over.

The most common adverse effects of risperidone are insomnia, anxiety, agitation, sedation, dizziness, rhinitis, hypotension, weight gain, and menstrual disturbances (9). Risperidone does not produce granulocytopenia or agranulocytosis, so no weekly monitoring is required, as is the case with clozapine. For this reason, it may be more acceptable to patients, all other aspects being equal. It may be expected to cause neuroleptic-malignant syndrome and galactorrhea.


The usual starting dose of risperidone is 1 mg twice daily, increasing to 2 mg twice daily and 3 mg twice daily over the next 2 days. Thus, it can be titrated to optimal therapeutic dose more rapidly than clozapine. The mean optimal dose appears to be 6 mg/day but higher doses may be needed to control positive symptoms in some patients. Since higher doses would increase the risk of EPS and presumably TD, lower dosages should be tried for adequate periods of time. If higher doses are needed acutely, lower doses should be substituted for maintenance treatment. Risperidone, like clozapine, should usually be given without concomitant typical neuroleptic drugs to avoid EPS. However, in a recent study, risperidone was superior to placebo as add-on-therapy to typical neuroleptic drugs in mentally retarded patients with persistent behavioral disturbances (72). This may reflect the value of 5-HT2 antagonism as a supplement to D2 receptor blockade.


Risperidone should prove to be useful as a first line antipsychotic agent. It could prove useful in a myriad of indications where typical neuroleptic drugs have been used: manic–depressive illness, psychotic depression, organic psychoses, especially in the elderly, L-dopa–induced psychoses, schizotypal and borderline psychoses, childhood psychoses, and so on. Its place in the armamentarium of antipsychotic drugs may well depend upon the ability of prescribers to obtain good control over psychopathology using doses of 4 to 6 and possibly 8 mg/day. If higher doses are frequently needed, the risk of clinically significant EPS, including TD, will be increased. The tendency of clinicians to keep increasing doses to achieve rapid response or to use typical neuroleptics concomitantly, will have to be altered for risperidone, if it is to fulfill its full potential as the first atypical first-line antipsychotic drug.


Remoxipride, a substituted benzamide, is a selective D2 receptor antagonist. Substituted benzamides such as sulpiride and amisulpride have been very widely used in Europe and Asia. Remoxipride is the first of this interesting class of drugs to be developed for use in North America. This class of compounds has been considered to have advantages with regard to treating negative symptoms and EPS, although the clinical trials to support this have not had the rigor to unequivocally establish these claims. A detailed review of the pharmacodynamic, pharmacokinetic, and early clinical literature is available (78). Animal studies indicate that doses required to block amphetamine-induced locomotor activity are significantly lower than those that produce catalepsy, indicating a potentially atypical profile. It may be expected to reduce mesolimbic dopaminergic activity while sparing the striatum (56). Because of its lack of effect on muscarinic, adrenergic, and histaminergic receptors (56), it produces few side effects. Patients often feel they are not taking any antipsychotic drug. It has a potent affinity for the sigma receptor, but it is unclear whether it is a sigma agonist or antagonist. The contribution of this property to its clinical profile is unknown (28).


The pharmacokinetics of remoxipride have been summarized (77). It is rapidly absorbed and has good bioavailability, with a plasma half-life of 4 to 7 hr. It has no active metabolites and is approximately 80% bound to plasma proteins. There are, as yet, no known drug interactions, but further study is required.

Clinical Studies

Remoxipride has been extensively studied in large double-blind, multicenter trials in comparison with haloperidol. Lewander et al. (32) provided a combined analysis of nine trials, involving 667 remoxipride- and 437 haloperidol-treated patients. After 4 to 6 weeks of treatment, 55% to 60% of patients in both groups were rated as much or very much improved. Haloperidol and remoxipride produced similar effects on positive and negative symptoms, with approximately 60% and 30% improvement in each class of symptoms, using an intention to treat analysis (32).

There is limited data on the usefulness of remoxipride in neuroleptic-resistant patients. In an open long-term trial of remoxipride in 45 chronic, neuroleptic-resistant schizophrenic patients, there was an 80% drop-out rate despite the absence of EPS. Ten patients (22%) showed a 30% or greater decrease in their total BPRS scores. These patients had relatively low mean BPRS scores at baseline: 18.4. Only 5 out of the 45 patients were considered very much improved (75). The clinically effective dose of remoxipride appears to be between 120 and 600 mg/ day (32).

Side Effects

In the composite summary of side effect data from nine multicenter trials, remoxipride appeared to have clear advantages over haloperidol with regard to EPS (32). This was especially true for interference with gait, elbow rigidity, fixation of position, head dropping, tremor, and salivation. However, 20% of remoxipride-treated patients required anticholinergic drugs, indicating that it does have significant EPS liability for vulnerable individuals. There is no evidence to suggest that remoxipride would not induce TD if given for prolonged periods. Of remoxipride-treated patients, 10% to 30% report drowsiness, tiredness, and difficulty concentrating, significantly less than haloperidol but not dramatically so.

Following the introduction of remoxipride in Europe, a significant number of cases of aplastic anemia were reported. The rate may be as high as 1 in 10,000. For this reason, the further use of remoxipride has been halted until the incidence, cause, and possible prevention of this side effect can be further studied. It is not clear if it will ever again be available without major restrictions and monitoring.


Remoxipride appears to be an effective antipsychotic drug that should be particularly useful in treating patients who experience significant EPS and sedation with standard neuroleptic drugs. Because of its ability to cause aplastic anemia, it may not be available for general use. Other substituted benzamides such as amisulpride, should be studied as possible substitutes.


There are now numerous drugs in advanced stages of clinical testing as potential atypical antipsychotic drugs. These are listed in Table 1 and briefly commented on here. The major focus of drugs in development is on 5-HT2A antagonists, based on the hypothesis that this property is a critical facet of the pharmacology of clozapine and risperidone. Melperone, a butyrophenone, which has been used in Europe for two decades, produces low EPS and has a definitely reduced liability to cause TD (48). It has also been found to be a relatively potent 5-HT2A antagonist compared to its D2 antagonist effects. Its absolute affinity for the 5-HT2A receptor is 30 nM (49). A large number of other antipsychotic drugs with a similar 5-HT2/D2 profile have been identified. These include Zaprisidone (65), olanzepine (53), Organon 5222 (66), seroquel (64), sertindole (67), and SM-9018 (25). In general, these compounds produce low catalepsy at doses that block amphetamine- or apomorphine-induced locomotor activity and the conditioned-avoidance response, and they have selective effects on the mesolimbic DA neurons, sparing the nigrostriatal system. Because they differ in relative potency as antagonists of a variety of other key receptors (e.g., D1, D3, D4, 5-HT1A, 5-HT2C, 5-HT3, 5-HT6, 5-HT7, M1, a1- and a2-adrenergic receptors), it is likely that these drugs will have significantly different side effect and efficacy profiles. Some may be effective in treating negative symptoms and improving cognitive dysfunction, others may not, and so on. Meticulous clinical research will be required to delineate the spectrum of action of these compounds. Olanzapine has recently been shown to produce significantly fewer EPS and to produce significantly greater decreases in positive and negative symptoms than haloperidol in schizophrenic patients in an acute exacerbation.

Two drugs related to this class are of special interest: amperozide (7, 15) and MDL 100,937 (70). Both are potent 5-HT2A antagonists but have no apparent D2-receptor blocking effects in vivo and yet have the preclinical or clinical profile, or both, of atypical antipsychotic drugs. Another drug that is a specific 5-HT2A blocker, ritanserin, has also been reported to be effective in treating positive and negative symptoms in schizophrenia. These results suggest that 5-HT2A antagonism by itself, may be a sufficient means of treating the full spectrum of symptoms in some types of schizophrenia.

There is considerable behavioral, biochemical, and electrophysiological evidence that 5-HT3 antagonists may be effective antipsychotic drugs (57), but clinical evidence is lacking to support this strategy (17). Blockade of 5-HT3 would be expected to decrease the release of DA, so this strategy is another variation on the theme of decreasing dopaminergic activity to alleviate psychosis. Whether 5-HT6 or 5-HT7 antagonists will be useful to treat psychosis or ameliorate EPS is unknown (see Serotonin Receptor Subtypes and Ligands and Indoleamines: the Role of Serotonin in Clinical Disorders). There is also some interest in the hypothesis that 5-HT1A agonist properties, in conjunction with D2-receptor antagonism, may increase the effectiveness and decrease the EPS liability of an antipsychotic drug.

The D1 receptor has long been known to modulate D2 receptor function in complex ways (see Dopamine receptors: Clical Corrrelates ). D1-receptor blockade has been thought to be a potential way to achieve an antipsychotic effect with low EPS (39). There are various D1 antagonists in clinical development (e.g., SCH 39166, NO-01-0687, DOD 647). Early clinical results with SCH 39166 have been unimpressive. However, this compound may be a partial D1 agonist. Trials with full D1 antagonists are underway. Since the cloning of the D3 and D4 receptors, which are localized in the mesolimbic, cortical, or striatal areas (see Molecular Biology of the Dopamine Receptor Subtypes, Mesocorticolimbic Dopaminergic Neurons: Functional and Regulatory Roles, and Dopamine receptors: Clical Corrrelates), the possibility of developing antipsychotic drugs based upon antagonists of these receptors has excited great interest (68, 74). There is as yet no preclinical evidence or clinical trials to suggest that drugs that are specific antagonists of the D3 or D4 receptors would be advantageous. Partial DA agonists (i.e., drugs that can possibly enhance mesocortical dopaminergic activity by a direct agonist effect while blocking DA-induced stimulation of the mesolimbic system) are being developed. To date, none of the D2 partial agonists (e.g., SDZ 208-911, SDZ 208-912 and terguride) have been systematically studied in schizophrenia (70).

The strategy of administering DA autoreceptor agonists to suppress the synthesis and release of DA is being studied. Several such compounds have now been developed, including BHT-920, 3-PPP, CGS 15873A, pramipexole, PD 118717, SND 919, WAY-124866 (see ref. 62 for a review of these compounds). The initial clinical studies with these drugs suggest a weak effect on negative symptoms (5).

Finally, the sigma receptor has been implicated in psychosis, but this has been disputed (54). There have been several attempts to test sigma receptor antagonists as antipsychotics (e.g., BMY 14802) without success. A new sigma antagonist with 5-HT2–receptor blocking properties DUP 734 appears to be of particular interest based upon its ability to block the behavioral effects of phencyclidine (71).


It seems clear that antipsychotic drugs that produce fewer EPS than typical neuroleptic drugs may do so via a variety of pharmacological strategies. The key to clozapine's advantages for positive, negative, and cognitive symptoms and signs in schizophrenia is not yet fully known. New antipsychotics that can be shown to share these advantages of clozapine will help identify its key pharmacological properties. It is likely that a multiplicity of effects contribute to the end result.


The research reported was supported in part by a U.S. Public Health Service grant MH 41684, and grants from the Elisabeth Severance Prentiss, Milton Maltz, and John Pascal Sawyer Foundations. The author is the recipient of a U.S. Public Health Service Research Career Scientist Award MH 47808. The secretarial assistance of Ms. Lee Mason is greatly appreciated.

published 2000