Acute Treatment of Schizophrenia

William C. Wirshing,M.D., Donna Ames Wirshing, M.D., Stephen R. Marder, M.D., and Theodore Van Putten, M.D.

Schizophrenia is, for many so afflicted, a chronic relapsing and remitting condition. While a person rarely returns to full psychosocial functioning during the periods of remission, the often times dramatic worsening that punctuate the typical clinical course have been the focus of much of the pharmacologic treatment research. These deteriorations—also called relapses and exacerbations—define the target symptoms of acute treatment studies. Their prevention is the goal of the various maintenance strategies (see "Long Term Treatments" chapter).

The acute pharmacologic phase of schizophrenia treatment concerns the introduction or reintroduction of medication to alleviate (or at least palliate) an exacerbation of psychosis. These episodes are usually marked by an increase in positive symptoms, such as delusions, hallucinations, thought disorder, and agitation. However, an increase in negative symptoms, such as extreme withdrawal or mutism, can also occur. An episode may be rapid or insidious in onset, and the form and content of the symptoms may change from one exacerbation to the next. Since the most important aspects of schizophrenic psychopathology involve subjective experiences (e.g., delusions and hallucinations), the ability or willingness of the individual to describe these phenomena reliably may also vary over time. Thus, the symptomatic target of acute pharmacotherapy is clinically elusive and at times simply unquantifiable.

Conventional neuroleptic agents have, since the mid 1950s, proven to be the most consistently effective compounds in the treatment of acute and chronic schizophrenic patients. This efficacy, though, comes at the cost of a number of untoward neurologic side effects. Prominent among these are disturbances of the extrapyramidal system, including dystonia, tremor, akinesia, bradykinesia, rigidity, akathisia, and a variety of tardive dyskinetic (TD) syndromes. These side effects account for the notorious patient noncompliance and iatrogenic morbidity (11,14,28,29,39,54,59,60). Additionally, conventional neuroleptics are only partially effective at ameliorating the psychosis, which contributes to persistent disability, subjective distress, and family burden. Finally, a substantial minority of patients derive little if any benefit from drug treatment (9).

This chapter will review the efficacy of conventional antipsychotic agents, the utility of plasma level monitoring, and the use of adjunctive agents in treating unresponsive cases. A discussion of the unconventional compounds is presented in the chapter on "Atypical Antipsychotics."

Once appropriate diagnostic, neuromedical, and psychosocial evaluations have taken place, the major considerations in acute pharmacologic treatment are the choice of drug, its dosage, and the dosage escalation schedule. Subsequently, the pharmacologic treatment plan should involve the assessment of therapeutic efficacy and adverse effects, the need for further dosage adjustment, and adjunctive or alternative treatments in those patients who fail to respond.

Several different classes of antipsychotic drugs have been introduced over the last 35 years. With the exception of clozapine (26), there are no convincing data that any one drug or class of drug is more effective than any other. Differences do exist, but studies with appropriate methodology have not been conducted to demonstrate these differences. Most comparisons involve random assignment, parallel-group designs (see Fig. 1) that contrast one drug with another and demonstrate a lack of significant difference in overall response rate.

These results do not necessarily mean that a given individual would respond equally well to either drug. Such designs typically exclude subjects with a history of placebo response and unresponsiveness to neuroleptics. These exclusionary criteria also greatly limit the generalizability of the results to a standard clinical population.

The equal efficacy data across classes of neuroleptics apply to the primary effects of these agents but not to the secondary or side effects. Generally, when compared with the low-potency neuroleptics, the high-potency compounds like fluphenazine, haloperidol, and droperidol produce less sedation, fewer and less severe anticholinergic effects, and fewer cardiovascular effects (e.g., orthostatic hypotension), but this comes at the cost of producing more acute extrapyramidal side effects (e.g., akathisia, dystonia and drug-induced parkinsonism). The choice of which neuroleptic to use is generally made by considering which particular constellation of side effects would be least harmful or most beneficial to a given patient. Though no consensus is available, many feel that the side effect profile of the high-potency agents is easier to manage for the clinician and better tolerated by the patient. This clinical impression may account for the fact that the high-potency agents are prescribed at two to seven times the dose (in chlorpromazine [CPZ] equivalents) as the low potency agents (4). However, because the high-potency agents carry with them a higher incidence of extrapyramidal effects, adjunctive medication (anticholinergic, dopaminergic, antiadrenergic) is frequently required.

One notion that continues to be widespread is that sedating drugs, such as chlorpromazine, are more effective for agitated or highly excited patients than non-sedating drugs, such as fluphenazine or haloperidol. The latter, in turn, are supposedly more appropriate for withdrawn patients or those with psychomotor retardation. This hypothesis has never been objectively confirmed, however, and numerous studies suggest that high- and low-potency drugs are equally effective in both types of patients.

Despite years of clinical and research experience, we do not have definitive dose-response curves for antipsychotic drugs. However, recent research provides much guidance to clinicians. In the early stages of antipsychotic drug development, it became apparent that chlorpromazine doses below 400–600 mg/day were much less likely to prove superior to placebo than doses above that range. Subsequently, particularly in the 1970s, there was considerable interest in exploring the upper ranges of tolerated doses to determine if such doses might produce any added benefit, either in terms of rapidity of therapeutic response or the ultimate degree of improvement. Studies comparing high-dose (defined as > 2,000 mg chlorpromazine equivalents) to standard-dose treatment showed no statistically significant advantage for the high dose (12,13,15,42,48,50,65). These findings do not exclude the possibility that some patients may benefit from higher than usual dosages, but such patients appear to be in the minority, and better means are needed to identify those individuals who might be appropriate candidates for high-dose treatment.

As Reardon et al. (49) have shown, an increase in the use of high dosages of high-potency neuroleptics occurred during the late 1970s and 1980s, despite the lack of clinical research data supporting such use. These high-dose regimens may have been the result of increased pressure to treat patients rapidly, the increasing acuity and severity of those being hospitalized, the belief by many clinicians that high doses of high-potency drugs are well tolerated, and the profound lack of treatment options for the patient who is unresponsive to neuroleptics. As a result of these trends, several studies in recent years have focused on clarifying the benefit/risk ratios of different neuroleptic dosages.

Levinson et al. (34) studied 53 patients with acute exacerbations of schizophrenic, schizoaffective (mainly schizophrenic), and other nonaffective psychoses. Patients were randomly assigned to fixed-dose, double-blind treatment with either 10, 20, or 30 mg/day of oral fluphenazine. After 24–28 days of treatment, improvement in the sample as a whole was unrelated to dosage. However, among patients who showed a 40% or greater improvement in positive symptoms, dosages of 0.3 mg/kg/day produced the greatest clinical improvement but also had a higher incidence of extrapyramidal side effects (EPS). The authors suggested that a linear relationship between fluphenazine dosage and clinical response exists among patients who respond to a certain degree, and that the nonresponder group may include many patients in whom dose is not a factor because they would be unresponsive to any of the dosages studied. The authors concluded that while the best clinical response was seen at dosages of 0.3 mg/kg/day, the increase in adverse effects was such that they would recommend daily dosages at the lower end of the 0.2–0.3 mg/kg range. They also found that the presence of akathisia during the study (regardless of whether or not it was treated) also predicted poor response during the four-week trial. Antiparkinsonian drugs were given as needed, but not prophylactically.

Van Putten et al. (62) reported on the treatment of 80 newly admitted men with schizophrenia who were assigned openly, by cohort, to receive 5, 10, or 20 mg/day of haloperidol for four weeks. Patients with a history of nonresponse to neuroleptic drugs were excluded, and patients with a history of severe dystonic reactions (28%) were given prophylactic benztropine mesylate (2 mg b.i.d.) After seven days of treatment, the proportion of patients who remained in the study and were described as "much improved" for the 5-, 10-, and 20-mg doses were 6%, 33%, and 47%, respectively. Although the 20 mg dose appeared superior in efficacy during the first two weeks, this group subsequently experienced a worsening in emotional withdrawal and psychomotor retardation, as well as a higher incidence of akinesia and akathisia. In addition, the 20 mg/day group had a 35% dropout rate (leaving hospital against medical advice) in comparison to only 5% each for the 5- and 10-mg dose groups. The investigators concluded that 20 mg may be more effective for controlling psychoses in the first week or two, but a higher incidence of adverse effects subsequently undermines this initial benefit. It is important to note that this was an open study (investigators were not blind to dosage), and it is possible that the prophylactic or early use of antiparkinsonian agents or propranolol to treat akathisia may have improved the therapeutic index.

Rifkin et al. (50) randomly assigned 87 newly admitted patients meeting Research Diagnostic Criteria (RDC; ref. 57) for schizophrenia to receive 10, 30, or 80 mg/day of oral haloperidol on a double-blind basis for six weeks. All subjects were given benztropine mesylate (2 mg t.i.d.). Among these patients, 93% had DSM-III (1) diagnoses of schizophrenia and 7% had a diagnosis of schizophreniform disorder. Although 22% of the subjects dropped out, no difference in dropout rate was observed among the treatment groups. Nor were there any significant differences between the treatment groups, either in terms of clinical response or in the occurrence of EPS. Thus, these investigators found no advantage to treating patients with haloperidol dosages >10 mg/day, but they also did not find a significant increase in EPS among patients treated with higher dosages and prophylactic antiparkinsonian medication.

There is a considerable degree of consistency in these studies, despite differences in methodology and patient populations. It would appear that there are no significant advantages to using dosages of haloperidol or fluphenazine >10–20 mg/day for acute treatment; even dosages of 20 mg may be associated with a substantial number of adverse neurologic effects if prophylactic antiparkinsonian medication is not used.

McEvoy et al. (42) used the "neuroleptic threshold" to determine optimal dosage for neuroleptic treatment of patients with acute schizophrenia. This involves a hypothesis first proposed by Haase (21) suggesting that the lowest neuroleptic dosages on which patients develop slight increases in rigidity are also the lowest dosages at which the patients will attain maximum therapeutic benefit.

Of the 106 patients who participated, 25 had schizoaffective disorder and 32 had no prior exposure to neuroleptics. On day 1 of the protocol, 2 mg of oral haloperidol was given and the daily dosage was subsequently increased by 2 mg every other day until the neuroleptic threshold was crossed (i.e., rigidity increased from baseline) or a dosage of 10 mg/day was reached. Once the neuroleptic threshold was reached within the first 10–12 days (very few patients failed to show an increase in rigidity on 10 mg/day), the dosage was fixed and the patient was treated in open-label fashion for 14 days. After 24 days, patients were randomly assigned, in a double-blind manner, to either continue at their neuroleptic threshold dosage or to have the dosage increased 2–10 fold. The average "high" dosage between days 24 and 38 was 11.6 + 4.7 mg/day versus 3.4 + 2.3 mg/ day for those continuing at their neuroleptic threshold dosage. Neuroleptic-naive patients crossed the threshold at a significantly lower average dosage (2.1 mg/day) than those who had been previously treated (4.3 mg/day). After 24 days (14 at the neuroleptic threshold dosage), 54% of patients were considered responders. After the double-blind comparison was completed at day 38, 42% of those who hadn't responded at day 24 had become responders, but there was no difference between those remaining on the neuroleptic threshold dosage and those randomized to a higher dosage. The only therapeutic measures on which the higher dosage was superior were ratings of hostility and suspiciousness. Unlike the Rifkin et al. study described above (50), higher dosages were associated with more EPS. In addition, the authors reported a significantly poorer response rate in those patients who had been actively psychotic for more than six months before treatment was initiated, compared with those having shorter periods of psychoses.

McEvoy's findings (38) differed from those of Levinson et al. (34) and Van Putten et al. (61), who found clear therapeutic advantages for the 10-mg dose, in comparison with lower doses. The inclusion of more first-episode and drug-naive patients in the McEvoy et al. (40) study could account for this, if one assumes that such patients are initially more sensitive to both the therapeutic and neurotoxic side effects of haloperidol. Monitoring for such subtle signs of neurotoxicity requires careful scrutiny and much clinical experience. Thus, the extent to which these findings are generalizable to routine clinical practice remains to be established.

Taken together, these results build a strong case that dosages greater than 15–20 mg/day of haloperidol or fluphenazine should not be the first-line treatment in patients who are judged to be capable of responding (i.e., those without an established history of neuroleptic refractoriness). It is also clear that neuroleptic side effects such as akathisia and akinesia are serious clinical problems even with dosages in this range, and efforts to prevent and treat them should be a high priority for clinicians.

Plasma level monitoring for antipsychotic agents has been of decidedly limited utility in both clinical and research settings. Among the reasons for this is that the antipsychotic drugs are highly lipophilic (i.e., little of the drug actually "resides" in the plasma space) and strongly protein bound (i.e., not bioavailable to cross the blood brain barrier). In addition, the correlation of dose to plasma level is generally low (intersubject variability is high), and clinical effect typically lags steady-state plasma levels by days or weeks. Some recent studies focusing on the relationship between plasma level and clinical response have, however, helped to characterize the potential usefulness and underscore the limitations of plasma level measurement of antipsychotics.

Early studies focused on drugs such as chlorpromazine that follow complex metabolic pathways. Plasma level measurements for these drugs are problematic, since some of the antipsychotic activity may be due to metabolites of the drug. Thus, drug selection and methodological errors may explain why early studies failed to demonstrate a reliable relationship.

More recent studies have focused on drugs other than chlorpromazine and have had more promising results. These studies have also employed improved methodology, including the use of fixed dosages. Haloperidol has received the most attention in this context, this is partially due to the fact that this drug has only a single important metabolite (reduced haloperidol), which may not have significant antipsychotic activity. Five studies found a "therapeutic window" relationship between plasma levels and clinical response, while five other studies did not. In most cases, failure to find a relationship may be explained by methodologic shortcomings, such as the use of patients with a history of poor drug response or the use of doses that were either too high or too low. At the same time, those studies that reported a poor response at higher blood levels may reflect an increase in adverse effects rather than a true decrease in efficacy.

The findings of Van Putten et al. (61) indicated a curvilinear relationship between plasma haloperidol levels (averaged during a fixed-dose treatment period) and changes in psychosis based on the Brief Psychiatric Rating Scale (BPRS; refs. 44,67). Patients demonstrated the most improvement when their plasma levels were between 5 and 12 ng/mL. Patients with levels above 12 ng/mL also improved as a group, suggesting that some patients tolerate these higher levels. However, when relative nonresponders had their plasma levels increased above 12 ng/mL, they failed to improve; some actually worsened. Volavka et al. (64) randomized 176 acutely ill schizophrenics to one of three plasma ranges of haloperidol: low (2–13 ng/mL), medium (13.1–24 ng/mL), and high (24.1–35 ng/mL). This innovative design permitted clinicians to evaluate the usefulness of targeting a particular plasma concentration. Overall, the three groups had approximately the same rates of response, although there was a suggestion that higher drug levels were associated with less improvement. However, the low plasma level range overlapped with what others consider the optimal range (e.g., 2–12 ng/mL; ref. 63). The findings of this study seem to indicate that there is no advantage to raising haloperidol levels above this "low" plasma range.

Given the array of studies and their varying results, it is understandable that no consensus exists as to whether or not plasma levels of antipsychotics should be monitored by clinicians. However, some conclusions may reasonably be drawn from an evaluation of the most recent generation of studies. Little is probably to be gained by monitoring plasma concentrations on a routine basis, since a high proportion of patients will respond when they are prescribed moderate doses of antipsychotics. On the other hand, a plasma level may provide useful information in the following circumstances: 1) when patients fail to respond to what is usually an adequate dose; 2) when it is difficult for the clinician to discriminate drug side effects—particularly akathisia or akinesia—from symptoms of schizophrenia such as agitation or negative impairments (i.e., a high blood level might be associated with increased adverse effects); 3) when antipsychotic drugs are combined with other drugs that may affect their pharmacokinetics, such as fluoxetine, beta blockers, cimetidine, barbiturates, and carbamazepine; 4) in the very young, the elderly, and the medically compromised—groups in which the pharmacokinetics of neuroleptics may be significantly altered; 5) when noncompliance or poor compliance is suspected; 6) when compliance is compelled by the legal system.

The bulk of the anticipated treatment-emergent toxicities from conventional compounds can be coarsely categorized into three general areas: extrapyramidal, cardiovascular, and anticholinergic. As already mentioned, the extrapyramidal toxicities are highly correlated with the drug's affinity for the D2 receptor and include the usual list of EPS. Embedded within this well known and generally well recognized constellation of neurotoxicities are more subtle impacts on mood and subjective toleration. The dysphoria induced by these agents, like the other EPS, is probably directly linked to D2 affinity characteristics and accounts, at least in part, for the markedly poor medication compliance in patients with schizophrenia (64). Although there is no consensus, most clinicians use anticholinergics prophylactically when instituting antipsychotic pharmacotherapy with agents that are equal to or greater than thiothixene in D2 affinity. Such a strategy minimizes the emergence of most dystonias, attenuates the akathisia and dysphoria, and does not add to the antimuscarinic load inherent with the low-potency compounds. Since geriatric subjects cannot tolerate anticholinergic medications (see below and Maintenence Drug Treatment for Schizophrenia), and because they are at lower risk for neuroleptic-induced dystonia and akathisia, anticholinergics are used only to treat, not prevent, the neurotoxicities.

The cardiovascular toxicities (e.g., tachycardia and orthostatic hypotension) are thought to be due to the a1 and M1 affinities. These properties are generally linked and inversely proportional to the D2 affinity, thus, they are more of a problem with the low-potency drugs. In subjects with anticipated intolerance of this toxic stress (e.g., the elderly), low-potency compounds are best either avoided or titrated slowly, as accommodation does occasionally occur. When clinically important tachycardia (generally >115 beats per minute) persists, it can often be managed with b1 blockers (e.g., atenolol). Orthostasis, though, should be monitored carefully when combining b1 and a1 blockers, as syncope may ensue. Orthostasis can usually be managed with a combination of dosage lowering, fluid loading, and (sometimes) the addition of pro-alpha agents (e.g., ephedrine). As alpha agonists directly increase heart rate, they may worsen the tachycardia when combined with strongly antimuscarinic agents (e.g., chlorpromazine, thioridazine, clozapine, etc.). However, when the orthostasis is due exclusively to a1 blockade (as in the case of risperidone which lacks M1 affinity; see below and Electroconvulsive Therapy), ephedrine may actually reduce the reflex tachycardia.

Ventricular arrhythmias have been associated with virtually all antipsychotic compounds, with thioridazine historically having the most notorious reputation. These arrhythmias, which are fortunately rare but unfortunately unpredictable, are thought to be secondary to prolongation of the QT interval, which then can result in a torsades de pointes tachyarrhythmia. Although speculative, this toxicity is probably due to a combination of a1 and H1 blockade. Thus, all of the newer drugs (risperidone, olanzapine, and sertindole) will predictably share this toxic liability with their conventional counterparts.

The anticholinergic toxicities include constipation, urinary retention, xerostomia, and accommodation disturbances (mostly in young patients). Less well known are the long-term embarrassment of deficits in neurocognition (e.g., short- and long-term memory, reaction time, attention, concentration, etc.). A growing body of evidence suggests that functional outcome in patients with schizophrenia is correlated more with neurocognitive abilities than with either positive or negative symptoms (see Long-Term Treatment of Mood Disorders). It is therefore a prudent clinical goal to minimize the overall anticholinergic load administered to patients—particularly the elderly (see Maintenence Drug Treatment for Schizophrenia).

The majority of controlled clinical trials have reported that 10–20% of schizophrenics derive little benefit from typical neuroleptic drug therapy (10). While the near future holds the promise of providing clinicians and researchers with the pharmacologic tools to safely and effectively treat this recalcitrant population (see the chapter on "Atypical Antipsychotics"), the present state-of-the-art is largely one of empirical trial and error.

Arguably the most common clinical choice for the treatment-resistant patient is high-dose neuroleptic therapy. The literature has a number of anecdotal (16,46,51) and controlled (24) reports supporting the use of very large doses (up to 60,000 mg/day of CPZ equivalents) in a treatment-resistant population. However, the study by Quitkin et al. (48) compared two fixed doses of fluphenazine (1,200 mg vs. 30 mg/day) in a double-blind manner in a nonchronic but treatment-resistant group of patients. The results favored the standard dose over the "megadose." Thus, while experience would support a high-dose trial for the treatment-resistant patient, it would also predict little benefit to such an approach.

Lithium has been used for over two decades to treat the symptoms of bipolar illness effectively. When combined with neuroleptics, lithium has also been reported to benefit patients with excited schizoaffective illness (5) and schizophreniform illness (22). Small et al. (56) showed that combining lithium and neuroleptics in a chronically hospitalized and relatively treatment-refractory population somewhat reduced symptomatology. It is therefore reasonable to try lithium in the neuroleptic-refractory patient, but, as with high-dose therapy, little optimism is warranted based on published data.

After Atsmon and colleagues (3) anecdotally reported that adjunctive high-dose propranolol positively influenced acute schizophrenia, a number of controlled studies addressed its utility as an adjunctive agent in treatment-refractory patients. Most (36,47,68) but not all (41) reported improvement with the addition of propranolol (400–2,000 mg/day) to standard neuroleptic regimens. Taken together, these results hint that high-dose propranolol might be a useful adjunctive agent, but they do little to guide the clinician in the choice of a target dose. A dose of 1,000 mg is probably a reasonable middle-ground choice (52), but support for even higher doses (up to 2,000 mg) can be found (36). Propranolol, like the antidepressants, elevates neuroleptic (and metabolite) levels (45), so care should be exercised in monitoring for an increase in neuroleptic-induced neurotoxic or endocrinologic side effects.

When used alone, carbamazepine has little to recommend it for stable but refractory schizophrenics. There is even some suggestion that it may destabilize some of these patients (58). However, some evidence from controlled studies indicates that CBZ, when combined with neuroleptics, may have benefits over a neuroleptic alone in "excited psychoses", including schizophrenia (33). Additional anecdotal reports and uncontrolled studies have indicated that it may be of adjunctive utility in schizophrenics with evidence of violence (37) or temporal lobe electroencephalographic (EEG) abnormalities (20,43). Others, though, found no benefit or even some worsening in non-excited but refractory schizophrenics (32). Kidron and colleagues (32) hypothesized that the measured reduction in neuroleptic plasma levels caused by the addition of carbamazepine (presumably through induction of hepatic metabolic enzymes) caused this clinical worsening.

So where does all this leave the clinician considering adjunctive carbamazepine for the treatment-refractory patient? It is probably reasonable to try it (at typical anticonvulsant levels) in refractory subjects with either known EEG abnormalities or with violent clinical manifestations. In addition to the usual hematologic, hepatic, and dermatologic concerns one has when using carbamazepine, the clinician must also be alert to the possibility that it may necessitate increasing the neuroleptic dose above baseline levels.

The literature on the use of benzodiazepines in schizophrenia is inconsistent, but it is somewhat skewed toward a negative or null effect. Some investigators have been mildly encouraging (25,30,35), but others have been frankly negative (19,23,27). Karson and colleagues (27) not only reported a lack of efficacy for adjunctive clonazepam but described the new development of violent behaviors in four of 13 patients during treatment. Thus, benzodiazepines should probably be reserved for those cases that fail all other adjunctive modalities or in whom clear anxious symptoms predominate.

Some investigators have reasoned that opiate agonists may have antipsychotic efficacy (18,31). Brizer et al. (7) used this reasoning to conduct a double-blind, single crossover, controlled study that compared methadone (25–40 mg/day) to a placebo as adjunctive compounds with neuroleptics in a group of seven treatment-refractory schizophrenics. The results indicated that methadone produced clinically modest but statistically significant improvement. The authors also stated that there were no difficulties getting these subjects off the methadone. While encouraging, such limited results cannot be extrapolated to routine clinical practice. Except in the most desperate and wretched of treatment-resistant cases, the use of adjunctive opiates cannot be justified.

Siris (55) continued his excellent clinical investigations into the etiology and treatment of post-psychotic depression and negative schizophrenic symptoms. His group reported on the effects of adjunctive imipramine in neuroleptic-stabilized schizophrenics who met criteria for both post psychotic depression and negative symptoms. Importantly, the protocol excluded any subjects whose symptom complex responded to a week's trial of adjunctive anticholinergics. This was an attempt to reduce the contribution and confounding influence of a neuroleptic-induced "akinetic" syndrome to the measured outcome. Even though the groups were small (17 placebo and 10 imipramine patients), the results showed that both depressive and negative symptoms improved together in the imipramine-treated group. This study is somewhat important from a clinical perspective, in that it provides the clinician with some medication options in this traditionally resistant group of schizophrenics. From the research point of view, it demonstrates that it is only through methodical pharmacologic probing at the junction of depression, negative symptoms, and neuroleptic toxicity that we will be able to understand and treat the contribution of each component.

Adjunctive specific serotonin reuptake inhibitors (SSRIs) can be beneficial to schizophrenic patients (2,17). Also, the addition of the 5-HT1A agonist buspirone has been of some benefit to small numbers of patients (2,8). Considering that serotonergic antagonism is among the explanations posited for clozapine's enhanced efficacy (see the chapter on "Atypical Antipsychotics"), it is theoretically curious that adjunctive putative serotonergic enhancing agents would improve some schizophrenic symptoms. Fluoxetine (or other SSRIs) may result in down-regulation of post-synaptic serotonin receptors, thus causing the overall effect of the medication to be similar to serotonin blockade. Alternatively (or perhaps additionally), it may work by increasing plasma levels of the neuroleptic through competitive metabolism (2,17). Higher levels of haloperidol, however, are not clearly correlated with good clinical response and indeed may be associated with poorer response (62). The fact that some patients will respond to drugs that have opposite effects on serotonin function underscores the need to develop better ways to select patients for one type of drug treatment or another. It may, for example, be that patients with schizophrenia who have obsessive-compulsive and depressive tendencies may benefit more from a trial of combined therapy with a conventional neuroleptic and an SSRI. Further research, particularly double-blind studies with SSRIs, is needed to substantiate the efficacy of these agents in schizophrenia treatment.

Although ECT is not as effective as medication across the range of schizophrenia patients, it has been shown to be of some value (53), and its relative merits n refractory patients deserves further study. Those patients with illness duration >6 months, significant affective symptoms, or catatonia are the ones most likely to benefit from ECT.

The two main contenders vying for a position along side of clozapine on the "atypical" shelf are risperidone and olanzapine. Risperidone is a benzisoxazole derivative that exhibits potent central antagonism of both serotonin (5-HT2A) and D2 receptors. It has also demonstrated substantial antagonism of a1, a2, and histamine (H1) receptors, but it is virtually devoid of activity at M1 receptors. Preclinical animal experimentation has indicated that, while it is slightly less potent than haloperidol as a D2 antagonist (canine emesis model), it is several times less potent than haloperidol at inducing catalepsy. Taken together, these data predict that risperidone may have antipsychotic efficacy and reduced EPS liability in humans. The potent 5-HT2A antagonism might theoretically be of utility in ameliorating the negative symptoms of schizophrenia, mirroring clozapine's enhanced efficacy in this spectrum of symptoms (see the chapter on "Atypical Antipsychotics"). Initial clinical reports published in 1993 (6,38) hinted that the most effective dose of risperidone (6 mg) may be more effective than 20 mg of haloperidol in controlling acute psychotic symptoms, and that the drug had an EPS liability that was not significantly greater than that of placebo. Data from the multi-center North American study further indicated that risperidone had dose-related extrapyramidal liability that begins to develop at a daily dose of about 10 mg. Even more interesting is that its antipsychotic efficacy began to wane at or above that same dose. Since its worldwide approval in 1993, risperidone has proven itself to be an effective antipsychotic with lower, but still present, neurotoxicity than conventional medications. This enhanced safety and tolerability have made it extremely popular, especially in the United States.

Olanzapine is a chemical analog of clozapine with affinity for D2, 5-HT2A, 5-HT2C, D1, a1, and M1 receptors. In clinical trials, it is equivalent to haloperidol in its impact on positive symptoms and slightly more effective on negative symptoms. Unlike risperidone, it has a more conventional, linear dose-response curve. Studies, however, have been limited to doses of 20 mg. It is not known whether olanzapine will, like its conventional counterparts, plateau above this dose range. Most compellingly, it has substantially less EPS liability than conventional drugs and demonstrates less elevation of prolactin (an indirect measure of D2 affinity) than risperidone. Olanzapine has only recently been approved for use in the United States, but pre-marketing data predict that it will begin to replace the older and more toxic conventional compounds (see Electroconvulsive Therapy).

An incremental advance in our clinical experience has come within the past few years with some of the new generation of antipsychotics, as well as a fine tuning of our understanding of the safest ways to treat acutely psychotic patients with conventional agents. The increasing clinical availability of new and different antipsychotic drugs will undoubtedly fuel even greater advances along these fronts in the near future. The experience gained from designing and conducting experimental clinical protocols with conventional compounds will clearly be carried over to the newer agents in the future.

The authors thank the National Alliance for Research on Schizophrenia and Depression (NARSAD), the Department of Veterans Affairs, and the National Institute of Mental Health for supporting their research work and Britton Smith, B.A. for his assistance with the manuscript.

published 2000