Print Friendly 

Ira D. Glick, MD, Xiaojing He, MD, and John M. Davis, MD

Primary Psychiatry. 2006;13(12):51-58

Dr. Glick is professor and Dr. He is research assistant in the Department of Psychiatry and Behavioral Sciences at Stanford University School of Medicine in California. Dr. Davis is professor in the Department of Psychiatry at the University of Illinois in Chicago.

Disclosure: Dr. Glick is on the advisory boards of Janssen, Pfizer, Shire, and Solvay; on the speaker’s bureaus of AstraZeneca, Bristol-Myers Squibb, Janssen, and Pfizer; receives research support from Bristol-Myers Squibb, GlaxoSmithKline, the National Institute of Mental Health, Pfizer, Shire, and Solvay; and is a stockholder in Forest, Johnson & Johnson, and Pfizer. Drs. He and Davis report no affiliation with or financial interest in any organization that may pose a conflict of interest.

Please direct all correspondence to: Ira D. Glick, MD, Stanford University Medical Center, Mail Code 5723, PBS #2200, 300 Pasteur Dr, Stanford, CA 94305-5723; Tel: 650-723-3519; Fax: 650-723-2507; E-mail:


Schizophrenia is a heterogeneous disorder in terms of symptoms, severity, course, and treatment response. Despite evidence of superiority of the second-generation antipsychotics (SGAs) over the first-generation antipsychotics (FGAs), there are still some in the field who argue that FGAs in low doses are equal to SGAs in efficacy and are far less costly. This article examines data comparing typical versus atypical antipsychotics on efficacy, safety and tolerability, and clinical strategies and indications. Possible reasons for still using FGAs include better or equal efficacy; better or equal safety and side-effect profiles; occasional efficacy of combining an FGA and an SGA compared to using either alone; reluctance to change an effective regimen (ie, “if it ain’t broke, don’t fix it,” pleas of patients who are doing satisfactorily on an FGA and are fearful of switching to an SGA, especially when they have minimal side effects); and when cost of treatment is a significant concern. Most clinicians who have used both FGAs and SGAs over the long run prefer an SGA unless cost is the primary determinant.

Focus Points


• Conventional antipsychotics are often used instead of atypicals for cost reasons.

• Over the long run, atypical antipsychotics are superior to typicals for most efficacy and side-effects measures.

• Metabolic side effects are an important issue for both typical and atypical antipsychotics.

• Treatment with antipsychotics must be individualized.



In the year 2006, is there any place for the typical antipsychotics? Conventional antipsychotics (ie, neuroleptics) were the mainstay of antipsychotic treatment for >40 years, with the advantages of proven short- and long-term efficacy, multiple routes of administration, and the availability of depot formulations. Their primary disadvantages were their short- and long-term side effects, including extrapyramidal symptoms (EPS), tardive dyskinesia (TD), sedation, sexual dysfunction, cardiac effects, prolactin elevation, anticholinergic effects, and weight gain. Thioridazine and mesoridazine have recently been required by the Food and Drug Administration to include black box warnings regarding their potential for QT prolongation and potentially life-threatening cardiac arrhythmia, although reported cases of such adverse events (AEs) are rare. “It is little wonder that acceptance of these medications was limited by physicians and patients alike, and noncompliance was a major factor in symptom relapse.”1


Kane2 noted that since the introduction in the mid-1990s, second-generation atypical antipsychotics have become far more widely used in the United States than the first-generation, or conventional, antipsychotics:

Given the increased costs associated with these medications, attempts to delineate their potential advantages and disadvantages from a clinical, public health, and health economic standpoint are certainly important. When evaluating the effects of antipsychotic drugs in schizophrenia it is important to recognize the heterogeneity of the illness, in terms of symptom patterns, severity, course, and treatment response (both therapeutic and adverse). Patients with schizophrenia present with positive, negative, cognitive, and other dimensions of symptomatology and functional impairment. These symptoms can worsen over time and can become less medication responsive. Patient’s vulnerability to side effects or dosage requirements for optimum response can also vary over time.2

Given this heterogeny, despite strong evidence of superiority of the second-generation antipsychotics (SGAs) over the first-generation antipsychotics (FGAs), there are still some in the field who argue that FGAs are equal to SGAs in efficacy and are far less costly. Accroding to De Paulo,3 the combined results of the three National Institute of Mental Health-sponsored trials—the Systematic Treatment Enhancement Program for Bipolar Disorder, the Sequenced Treatment Alternatives to Relieve Depression studies, and the Clinical Antipsychotic Trials Intervention Effectiveness (CATIE) study for treatment of schizophrenia—underline the suggestion that modern pharmacologic treatments may be no more beneficial than older ones, despite their added cost. This is not a minor issue because the costs of new drugs are much more than the old.

Accordingly, this article examines the data comparing typical versus atypical antipsychotics on efficacy, safety and tolerability, and clinical strategies and indications.

Rationale for Use of FGAs

Although most clinicians around the world have switched to SGAs, Jefferson4 believes that clinicians should continue to be taught about FGAs for several reasons:

First, there are many patients who have both benefited from and tolerated the older antipsychotics and who have no need to switch to something new for the sake of newness. Second, treatment resistance and partial response are all too common in schizophrenia, which has led to the frequent use of combination therapies despite a paucity of research to justify this approach. Nonetheless, combinations of atypical and typical or two or more atypical antipsychotics are quite common (just as combinations of typical antipsychotics once were). Third, many would argue, albeit without research support, that a patient should not be defined as treatment resistant without having failed an adequate trial of at least one typical antipsychotic. Finally, while, on the whole, the atypicals are better tolerated, they are not without their own blemishes. For example, there is growing concern about the long-term use of at least some of the new-generation drugs.4

In summary, reasons to still use FGAs include better or at least equal efficacy; better or equal safety and side-effect profiles; that combining an FGA plus an SGA may be better than either alone; patient reluctance to change an effective regimen if they are doing satisfactorily with minimal side effects on an FGA; and that cost of treatment is a significant concern.

Data Comparing FGAs Versus SGAs on Efficacy

This section reviews efficacy in the acute and maintenance stages of treatment as well as for specific symptom domains. Overall, there has been strong evidence favoring some SGAs over FGAs at all phases of illness (Table 1).1,5 Two studies that challenge this conclusion are discussed below.


General Considerations

It is important in the use of medication to take into account both evidence from controlled clinical trials as well as the prescribing doctors’ experience with a particular patient in that particular practice. In evaluating evidence, special attention should be placed on the actual results of controlled, random assignment, double-blind clinical trials rather than the conclusions various authors can draw from the trials. The actual results are not influenced by industry bias (J Davis, N Chen, I Glick, unpublished data, October 2006). In other words, the double-blind does work.
The actual numerical results are not influenced by the drug company, although the conclusions the authors draw may be influenced. There are wide differences of conclusions drawn by various authors, but close examination of the actual results of the trials are mostly consistent. This contrasts with the narrative conclusions drawn which are absolutely inconsistent. Almost all the proponents from each drug company argue that their drug is better. Research supported by cost-conscious public agencies almost always favor the most inexpensive medications, the first-generation antipsychotics. This article shows what the actual results are and offers the authors’ opinion about what is clinically important. The authors of this article believe that the clinical framework for interpretations of results are pertinent. This article compares and contrasts studies that measure the acute effects of antipsychotics given to patients in an episode and studies on the long-term maintenance medication given to prevent relapse.

Acute studies start with an acute psychotic patient in a psychotic episode generally admitted to a hospital. The usual design is that a new drug, old comparison drug, or placebos are administered in random-assignment double-blind studies, and improvement is measured. The patients show much improvement the first week, but the rate of improvement begins to decrease and the improvement curve flattens out and slowly moves toward maximum improvement. Patients achieve perhaps half their improvement in the first week or so, although it takes 4–8 weeks or even longer to see all the improvement that will occur. The design of long-term maintenance trial is substantially different for patients who are recovered and then maintained on a new drug, old drug, or placebo for 6 months to 1 year. On placebo, approximately 10% of the patients relapse each month, so that it takes 7 months for 50% of the patients to relapse. There may be as large as a 10-fold difference between the rate of improvement of the acute patient to the rate of relapse of stabilized patient. Meta-analyses select studies by a certain criteria and then essentially average up the improvement on the new drug, old drug, or placebo, weighting improvements by the number of patients in each trial.

In long-term maintenance medication studies, patients start out well and then begin to relapse. One of the most common measure of improvement is time to relapse. Patients on placebo relapse at a rate of approximately 10% per month, while patients on active drugs may relapse at a rate of 1% to 3%. The usual index of relapse is hazard rate, the rate of relapse of new drug to standard drug. In most long-term clinical trials, generally patients are only studied on either a new or standard drug. This methodology is often referred to as “survival analysis.” In registrational trials, the focus is on keeping patients in the trials as long as possible. Patients are dropped only for good reasons in order to prove safety and efficacy to the regulatory body such as the FDA. Patients are kept in trial until relapse (at which point it would be unethical to keep a patient in the trail) or intolerable side effects occur. There are also always administrative reasons that lead to dropping out.

As mentioned, in maintenance studies, patients start out well and become sick. It is the mirror image of acute treatment studies where patients start out sick and become well. In maintenance studies, the usual index is months until relapse or in some cases months until all-cause discontinuation. All-cause discontinuation refers to patients who drop out of the trial for any reason, whether because they are dissatisfied with the efficacy of the drug, because of a side effect, or because of administrative reasons.

It is important to introduce the clinical concept that only some patients will have a given side effect. Since a drug may have a higher instance of one side effect and a lower instance of another side effect, or complete absence for that matter, it is often possible to find a drug which produces considerable improvement but without the given side effect that the patient had with a different drug. If a pharmacy was only going to stock one drug, they might want to choose the drug which produced maximum efficacy with minimal overall side effects; however, because drugs have different side effects, use of a drug with a given side effect in a patient who has that side effect can be avoided. It is quite meaningless to work with any concept which implies that all drugs have the same side effects in the context of the individual patient. For overall evaluations, it may be a useful abstraction; however, when applied to the individual patient, a drug should be chosen for which the patient has a good response without significant side effects. Not every patient has all side effects, ie, the drugs produce different frequencies of side effects and there is room for much individualization. Therefore, it is important to keep efficacy separate from side effects and the side effects separate from each other. This article presents available information on overall efficacy and on the prevalence of different side effects, to give the clinician the information needed to make intelligent choices. This article also identifies areas where there is no information from controlled studies.

Acute Treatment

Meta-analysis always compares new drugs against a standard. A convenient benchmark is that typicals are 0.6 effect-size units over placebo and clozapine is 0.5 effect-size units over typical antipsychotics. The olanzapine, risperidone, and amisulpride cluster of second-generation drugs are approximately 0.25 effect-size units over typical antipsychotics, or approximately one-third as efficacious as typical antipsychotics are over placebo.5 The authors of this article caution against being overconcrete in interpreting results from meta-analysis. A meta-analysis by Davis and colleagues5 found clozapine approximately .5 effect-size units better than typical antipsychotics, but typical antipsychotics are approximately .6 effect size-units over placebo. The olanzapine, risperidone, and amisulpride cluster of second-generation antipsychotics are approximately .25 effect sizes superior, and so would be one-third as efficacious over typical antipsychotics as typical antipsychotics are compared to placebo.5 Meta-analyses always provides a comparative judgment against a standard.

While the typical antipsychotic haloperidol has been considered the standard of care for acute positive symptoms of schizophrenia, several studies challenge that belief and suggest that the atypical antipsychotics olanzapine and risperidone are equally effective in managing acute psychosis and may further reduce the risk of relapse.6,7 One analysis of double-blind extensions of three acute studies reported a 1-year risk of relapse of 19.7% with olanzapine and 28% with haloperidol (P=.034).6 Another study comparing an oral regimen of risperidone plus lorazepam (benzodiazepine) with an intramuscular (IM) regimen of haloperidol plus lorazepam reported similar improvements in agitation and time to sedation, with similar tolerability and safety results.7 A study of ziprasidone IM reported mean improvements in the Brief Psychiatric Rating Scale (from 47.8 to 28.9) and Clinical Global Impressions scale that were maintained on days 4 and 5 of therapy, with no episodes of EPS, acute dystonia, or AEs reported.8

Multiple meta-analyses comparing typical agents with atypical antipsychotics in schizophrenia have shown that atypical antipsychotics are equivalent or more effective than typical agents in preventing relapses.9,10 There are several first-episode studies where patients are followed for a significant length of time (at 6 months, 1 year, and 2 years) on a second-generation drug versus a first-generation. While efficacy differences in the first few weeks are negligible, there is some evidence that there are substantial efficacy differences at 1 year. There is a suggestion that olanzapine prevents further enlargements of ventricles during this time period in comparison to haloperidol. There is evidence of increased cognitive function and better improvement over the long term.11 It is certainly clear that schizophrenia is not fully developed in the first episode. Patients have a greater degree of symptom improvement in the first episode than in years later in one of the many multiple relapses that these patients undergo. It would be premature to say at this point in time that it is proven that some second-generation antipsychotics prevent some of the progression, tissue loss, neurocognitive decline, and negative symptom development that develops after the first episode. The possibility that this is true should not be ruled out at this time, and in making decisions the clinician needs to consider the consequences that this possibility may be true or false.

Long-Term Treatment

Long-term therapy with both typical and atypical agents has been studied to determine the effects on reducing relapse rates. One placebo-controlled study demonstrated that treatment with olanzapine reduced relapse rates from 55% with placebo to 5.5% with olanzapine over a period of >200 days.12 More importantly, the classic (and first long-term) randomized, double-blind study by Csernansky and colleagues13 compared risperidone with haloperidol therapy, finding that those receiving risperidone had a significantly higher chance of remaining relapse-free after a follow-up period of >2 years (34% risk in the risperidone group versus 60% risk in the haloperidol group; P<.001), representing a 48% risk reduction with risperidone. Similarly, a retrospective analysis of olanzapine, risperidone, clozapine, and haloperidol over a period of at least 1 year reported that at 24 months, 87% of clozapine-treated patients and 66% of risperidone-treated patients remained unhospitalized, while the haloperidol patients relapsed at far higher rates.14

The CATIE trials were designed to be long-term trials.15 However, if either the doctor was dissatisfied with the improvement, the patients wanted to switch, or side effects occurred, the patients could be switched to one of two additional arms. Thus, if either the doctor or the patient gave up on a drug and/or did not like side effects, they had the option of continuing in the trial and being re-randomized. They were then given the choice of whether to be re-randomized, ie, where they may receive clozapine as one of the drugs, or they may decline to accept clozapine and be re-randomized with just atypical antipsychotics. After completion of this phase, they could still remain in the trail on any open medication. This is an important design difference because in the industry trials the pressures were to keep patients in the trial. In the CATIE study, patients could remain in a trial but go on to get a second drug. Furthermore, in the CATIE study, if the second drug did not work they could remain in a third phase of the study where they could receive any drug their doctor wanted to use. Some have called the choice to move from the first drug to the second as a desperation choice. Others have suggested that it is more of a wishful thinking trial, ie, the patient is not fully satisfied, wishfully thinks maybe another drug would be better, and elects to go on to the second phase and after that to the third phase. The desperation or wishful thinking interpretations are both correct as they use different terms to describe the same phenomenon of all-cause discontinuation, which was a broad measure for patients that switched under the incentives of the CATIE trial for various reasons. Therefore, the index of drug difference overall is a composite. The average patient in the CATIE study remained on olanzapine for 9.2 months in comparison to 3.5 for ziprasidone or 4.6 for quetiapine. As an index of which drug was more effective than the other, olanzapine was roughly twice as effective. These results can be analyzed as hazard rates, which is a comparison of one treatment to another treatment. If drugs had exactly the same relapse rate or discontinuous rate then the hazard rate would be 1. In the CATIE study, patients on ziprasidone were 1.6 times more likely to discontinue medication or 0.63 times as likely to remain on medication as olanzapine. In addition, a recently published study that focused on patients whose medication was changed for clinical reasons found no disadvantage across 1 year in using FGAs rather the SGAs.16 In this study, like the CATIE study, important design questions have been raised. Nevertheless, the findings of both studies have added fuel to the multi-million dollar cost/efficacy issue.

Results from these two studies should be put in context with the meta-analysis by Davis and colleagues.5 In order to compare acute efficacy versus maintenance efficacy, this meta-analysis shows consistent evidence that clozapine is much more efficacious than typical antipsychotics; some evidence that risperidone, olanzapine, and amisulpride are somewhat more efficacious than typical antipsychotics; and evidence that the majority (6 of 10) of atypical antipsychotics have approximately the same efficacy as typical antipsychotics. Obviously, some of the CATIE findings are not consistent with this analysis. Discussion of these methodic issues are outside the scope of this article and can be found in several recently published articles.17,18 Other recent articles, which attempt to resolve these discrepancies, also warrant discussion.19,20

Does adding psychotherapy differentially affect FGAs versus SGAs? Combination therapy with pharmacologic treatment and psychosocial therapy may further improve outcomes. A study by Marder and colleagues21 found that those receiving risperidone plus psychosocial skills training had the highest incidence of continuing therapy over a period of approximately 3 years compared to those receiving haloperidol plus psychosocial training (58% of patients receiving risperidone versus 40% of those receiving haloperidol, P=.19).21 Patients receiving risperidone also had significantly higher scores in terms of quality-of-life and social adjustment (P=.05 for both).17 In a similar study, clozapine-treated patients had a higher rate of participation in psychosocial rehabilitation treatment, were more likely to both verbally accept recommendations and to act on them, and had higher rates of reduced symptoms and improved quality-of-life at 12 months than the haloperidol group.22

Other Efficacy Considerations

Although limited controlled data are available with respect to cognitive function, preliminary findings suggest that atypical agents may be more efficacious in improving cognitive ability, including attention, memory, verbal fluency, and executive function, as compared with typical agents.23 Recently, the CATIE study found small, relatively equal improvements in both FGAs and SGAs on cognition over 18 months.24

Should an FGA be combinied with an SGA? The only data suggesting that this type of combination is better than one antipsychotic are anecdotal, ie, there are no controlled studies. Despite that, it seems that almost every clinician has at least one patient on two antipsychotics. The major clinical pearl is to (on a periodic but continual basis over the years) rechallenge this regimen, ie, slowly discontinue one medication and see if the patient’s symptoms exacerbate.

It has been suggested that SGAs seem falsely more efficacious than FGAs as a result of reduced efficacy due to use in the studies of a high-dose comparator, haloperidol.25 As mentioned earlier, Davis and colleagues5 performed a meta-analysis of randomized efficacy trials comparing SGAs and FGAs, comparisons between SGAs, a dose-response analysis of FGAs and SGAs, and an analysis of the effect on efficacy of an overly high dose of an FGA comparator. Some SGAs were found to be more efficacious than FGAs, and, therefore, SGAs are not a homogenous group. Rosenheck25 has shown that over the long run there is equal efficacy of haloperidol with a mean dose of 17 mg compared to olanzapine.

It is important to recognize that there are substantial methodologic differences between much of the service research studies, such as the one by Rosenheck,25 and that of controlled clinical psychopharmacologic trials conducted by the drug industry or academic trialists. Service research studies typically include data from patients randomized to the study drugs in the trial even though they stopped taking medications or switched to another medication. Their data are used whether or not the patients take study drugs. The usual clinical trial drops such patients out and only uses the data from when the patient was on study drugs. For example, a patient randomized to olanzapine but switched to a typical antipsychotic would still be counted in the trial data as if he or she received olanzapine, even though olanzapine might have only been taken in the first few months. A patient randomized to haloperidol but switched to clozapine would have the improvement on clozapine count in the haloperidol arm. This design tends to mute any differences between drugs. There is an important rationale for the economics of medical care for the use of this type of methodology, but the details of which are beyond the scope of this article. Although data from Rosenheck25 was included in the meta-analysis of long-term trials by Davis and colleagues,5 no evidence of equal efficacy was found.

The CATIE trial recently reported (in Phase I) that perphenazine had equal discontinuation rates as risperidone and quetiapine but more than olanzapine.15 However, the proportion of patients with EPS did not differ significantly among those who received FGAs and SGAs in that study. The study did note that despite this finding, more patients discontinued perphenazine than other medications due to EPS.

Safety and Tolerability of Antipsychotics

Do side-effect profiles between FGAs versus SGAs help the clinician to make the decision of whether to use FGAs? Each available antipsychotic regimen is accompanied by certain AE risks (Table 2).1,26 Clinicians must therefore monitor patients carefully and consider pre-existing comorbidities when selecting an appropriate treatment strategy. Recent analysis of discontinuation rates based on inability to tolerate AEs in the CATIE study showed that of those compared, risperidone had the lowest discontinuation rate (10%), followed by quetiapine and ziprasidone (15% each), perphenazine (16%), and olanzapine (19%).15 Although studies guide clinicians on expected rates of medication effectiveness, individual patients can be markedly different in their ability to tolerate particular medications with different side-effect profiles.

Although atypical antipsychotics demonstrate some degree of EPS and TD, it is clear from a review of clinical data that the risks of suffering these effects are more common with typical agents. Somnolence, while often seen with low-potency typical antipsychotic regimens, is also common in patients receiving quetiapine and clozapine.26 Another AE noted with clozapine is agranulocytosis or neutropenia, which increases the risk of infection and can be fatal if undetected.1 However, the risk is very low and rarely seen clinically. Orthostatic hypotension is a concern with low-potency typical agents but is less common with some but not all of the atypical antipsychotics.26 Likewise, FGAs do not appear to be safer than SGAs in the elderly.27

Weight gain and obesity resulting from antipsychotic therapy can affect long-term health, quality-of-life, and medication compliance, thereby increasing the risk of relapse. Related to long-term health and quality-of-life are an increased risk of heart disease, diabetes, low self-esteem, and early mortality.28-33 Atypical antipsychotics are more likely to cause weight gain than typical antipsychotics. Clozapine, olanzapine, and quetiapine are the agents that are most commonly associated with weight gain.26,28,29 Likewise, some typical antipsychotics, especially the low-potency agents thorazine and thioridazine, also may cause weight gain. Even high-potency FGAs can cause appreciable weight gain in young, thin, or first-episode patients. Weight gain is rapidly progressing over time, although the amount of weight gain per month decreases and eventually plateaus. Patients who remain on a weight-gaining drug for an average of 9 months will gain substantially more weight than those remaining on that drug for an average of 4 months, by virtue of time on that drug.

The potential risk of metabolic syndrome (the presence of two or more metabolic risk factors such as abdominal obesity, elevated blood pressure, insulin resistance, dyslipidemia, and type 2 diabetes) exist with antipsychotic therapy.33 In Phase I of the CATIE study, olanzapine had the greatest rate of weight gain and increased glycosylated hemoglobin, total cholesterol, and triglycerides, which are related to the potential development of metabolic syndrome.15 In terms of weight gain, 30% of olanzapine patients gained >7% of their baseline weight, compared to 7% with ziprasidone, 12% with perphenazine, 14% with risperidone, and 16% with quetiapine.15 Ziprasidone was associated with improvement in several metabolic categories, with a median 1 mg/dL and 7 mg/dL decrease in total cholesterol and triglycerides, respectively. Olanzapine had a 9% discontinuation rate due to metabolic effects and weight gain, compared to 4% with quetiapine, 3% with ziprasidone, 2% with risperidone, and 1% with perphenazine (P<.001).

New Strategies and Indications for Therapy

Adequate pharmacologic interventions complimented by psychosocial support, family counseling, and rehabilitation are critical to reduce long-term deterioration. Because every relapse reduces the level of cognitive function, compliance is a critical goal and is improved with these interventions.34 While most patients will tolerate some AEs to avoid the major effects of schizophrenic illness, it is important to emphasize treatment strategies that provide a minimal effective dose of life-long duration, with continuous rather than targeted treatment for all phases of the illness over a lifetime. The most important point to remember when selecting therapy is that treatment strategies for each patient must be individualized to attain the greatest benefits with minimal AEs.

Psychotherapy (both individual and family) and psychoeducation should always be combined with any pharmacologic regimen in schizophrenia. Although there is no conclusive evidence that a combination regimen of typical and atypical antipsychotics improves outcomes, some believe that there is a small population of patients who may improve with such therapy. Mood stabilizers and antidepressants are sometimes prescribed in the acute phase of treatment, but there is no evidence that these agents confer benefits in long-term management. Benzodiazepines can be used throughout the course of the illness to manage agitation. Acute IM forms of antipsychotic therapy are often needed in noncompliant patients during acute episodes. Many typical antipsychotics have acute IM preparations. Among the atypical antipsychotics, olanzapine, ziprasidone, and risperidone are currently available in IM forms; other IM formulations are also in development in an effort to increase therapeutic options in noncompliant patients. Finally, a subset of patients are unable to sustain compliance with outpatient oral medication. Long-acting IM forms of risperidone as well as two high-potency typical antipsychotics (haloperidol and fluphenazine) can provide effective alternatives for some of these patients.

Treatment regimens are sometimes switched for reasons of safety, tolerability, and efficacy. Although switching therapy may help in the short run, it does not usually improve patient status in the long run. If switching is necessary, clinicians should focus on target symptoms, such as improvements in mood or cognition, or a resolution of AEs, and document change (or its absence). In the absence of a written record, switching without record keeping serves very little use. Because most patients with schizophrenia only achieve a partial response, it is very common to find patients not only on many drugs, but also frequently switching from drug to drug. Unless some quantitative record is kept about therapeutic benefits or side effects of patients on one or another drug, there is no way to learn whether that drug was particularly beneficial or had particularly troublesome side effects except in the extremes, where a patient may remark that a given side effect was particularly noxious and possibly refuse to continue on the drug. More subtle benefits like a decrease in negative symptoms may not be readily communicated.

There is no evidence that a given antipsychotic may be particularly efficacious with one group of patients and less efficacious with a different group. It would be possible that a patient might respond uniquely to haloperidol but not to olanzapine, risperidone, quetiapine, etc; however, there is no concrete evidence as it has never been studied. It is certainly possible that one patient may respond well to haloperidol and a different patient to olanzapine. If this were true, there would be a rationale for experimentation which might identify a drug for which a patient is particularly responsive. As mentioned, it is known that some antipsychotics are somewhat more efficacious than others. More importantly, there are quite marked differences in side effects between drugs and there is a good likelihood of finding a drug which does not cause the particular trouble from a particular side effect. Conversely, it is important not to avoid a very efficacious drug because of worry about a side effect that the patient may not have.

In any case, schizophrenia is a disease which lasts throughout life and, therefore, even a small efficacy advantage would surely be quite clinically important over the course of many years. The hope would be that combining the most effective pharmacotherapy with assertive case management, cognitive-behavioral therapies, and vocational rehabilitation may restore patients to a better lifestyle so that they could live more normal and productive lives. There is good evidence that olanzapine and risperidone—to a modest degree—improve function, negative symptoms, cognitive symptoms, anxiety, and depression. It remains unclear if putting both the psychosocial rehabilitation and the optimal therapy together in the long run will lead to true rehabilitation. Many of the patients in assertive case management of vocational rehabilitation and other psychosocial programs are receiving SGAs. It is quite possible some of the good results achieved in these programs require as their basis some of these SGAs, although it is important to add that this cannot be proven either. Psychotherapy may help in the short-run, but it does not usually improve patient status in the long run.

Although schizophrenia has an important pharmacoeconomic impact on the healthcare system and atypical antipsychotics are more expensive than typical therapies, it is important to note that drug costs account for only approximately $5 billion of the estimated $62.7 billion spent in 2002 on the care of patients with schizophrenia.35 In fact, most retrospective analyses suggest that effective treatment with atypical antipsychotics reduces overall costs and improves quality-of-life. One naturalistic study convincingly suggests that quality-of-life is much better with atypical than typical antipsychotics.36 Even so, most governmental agencies often opt for lower cost when choosing an antipsychotic.


Although the data in the literature over the last decade suggest that SGAs are more advantageous than FGAs, for a minority of clinicians the issue is still not settled. For them, in those clinical situations where FGAs are used, the CATIE study provides a rationale for their use. Two recent commentaries by CATIE investigators succinctly summarize the issues for the field.37,38 Therefore, the most cogent reason at this point to use FGAs is cost. In the same vein, an FGA is better than no treatment at all.12 Secondly, for a minority of patients, an FGA may be useful at some point in their illness but usually not over a lifetime.39 By and large, most clinicians who have used both classes prefer to use SGAs. As one patient put it, the SGAs compared to FGAs are “lighter and airier; I do not feel as if I have a blanket on my brain. I can do logistics without making errors, am more on autopilot, and can cope without panic” (ID Glick, MD, personal communication, October 2003). Although a testimonial is not as scientific as a controlled study, this one describes the nuances of these two classes of antipsychotics, and, as such, it may be useful in guiding clinical practice. Needless to say, only further research comparing FGAs with SGAs will settle these issues. PP



1. Jibson MD, Glick ID, Tandon RT. Schizophrenia and other psychotic disorders. Focus. 2004;2:17-30.

2. Kane JM. Commentary on the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE). J Clin Psychiatry. 2006;67(5):831-832.

3. DePaulo JR Jr. Bipolar disorder treatment: an evidence-based reality check. Am J Psychiatry. 2006;163(2):175-176.

4. Jefferson JW. Old versus new medications: how much should be taught? Acad Psychiatry. 2005;29(2):162-166.

5. Davis JM, Chen N, Glick ID. A meta-analysis of the efficacy of second-generation antipsychotics. Arch Gen Psychiatry. 2003;60(6):553-564.

6. Tran PV, Dellva MA, Tollefson GD, Wentley AL, Beasley CM Jr. Oral olanzapine versus oral haloperidol in the maintenance treatment of schizophrenia and related psychoses. Br J Psychiatry. 1998;172:499-505.

7. Currier GW, Simpson GM. Risperidone liquid concentrate and oral lorazepam versus intramuscular haloperidol and intramuscular lorazepam for treatment of psychotic agitation. J Clin Psychiatry. 2001;62(3):153-157.

8. Brook S. A pilot study of intramuscular ziprasidone in the short-term treatment of patients with acute exacerbation of schizophrenia. Hum Psychopharmacol. 2000;15(7):521-524.

9. Leucht S, Pitschel-Walz G, Abraham D, Kissling W. Efficacy and extrapyramidal side-effects of the new antipsychotics olanzapine, quetiapine, risperidone, and sertindole compared to conventional antipsychotics and placebo. A meta-analysis of randomized controlled trials. Schizophr Res. 1999;35(1):51-68.

10. Leucht S, Barnes TR, Kissling W, Engel RR, Correll C, Kane JM. Relapse prevention in schizophrenia with new-generation antipsychotics: a systematic review and exploratory meta-analysis of randomized, controlled trials. Am J Psychiatry. 2003;160(7):1209-1222.

11. Lieberman JA, Tollefson GD, Charles C, et al. Antipsychotic drug effects on brain morphology in first-episode psychosis. Arch Gen Psychiatry. 2005;62(4):361-370.

12. Dellva MA, Tran P, Tollefson GD, Wentley AL, Beasley CM Jr. Standard olanzapine versus placebo and ineffective-dose olanzapine in the maintenance treatment of schizophrenia. Psychiatr Serv. 1997;48(12):1571-1577.

13. Csernansky JG, Mahmoud R, Brenner R, the Risperidone-USA-79 Study Group. A comparison of risperidone and haloperidol for the prevention of relapse in patients with schizophrenia. N Engl J Med. 2002;346(1):16-22.

14. Conley RR, Love RC, Kelly DL, Bartko JJ. Rehospitalization rates of patients recently discharged on a regimen of risperidone or clozapine. Am J Psychiatry. 1999;156(6):863-868.

15. Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Eng J Med. 2005;353(12):1209-1223.

16. Jones PB, Barnes TR, Davies L, et al. Randomized controlled trial of the effect on quality of life of second- vs first-generation antipsychotic drugs in schizophrenia: Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS 1). Arch Gen Psychiatry. 2006;63(10):1079-1087.

17. Tandon R, Nasrallah HA. Subjecting meta-analyses to closer scrutiny: little support for differential efficacy among second-generation antipsychotics at equivalent doses. Arch Gen Psychiatry. 2006;63(8):935-937.

18. Davis JM, Chen N, Glick I. Subjecting meta-analyses to closer scrutiny: little support for differential efficacy among second-generation antipsychotics at equivalent doses. Response. Arch Gen Psychiatry. 2006;63(8):938-940.

19. Moller HJ. Are the new antipsychotics no better than the classical neuroleptics? The problematic answer from the CATIE study. Eur Arch Psychiatry Clin Neurosci. 2005;255(6):371-372.

20. Duggan M. Do new prescription drugs pay for themselves? The case of second-generation antipsychotics. J Health Econ. 2005;24(1):1-31.

21. Marder SR, Glynn SM, Wirshing WC, et al. Maintenance treatment of schizophrenia with risperidone or haloperidol: 2-year outcomes. Am J Psychiatry. 2003;160(8):1405-1412.

22. Rosenheck R, Tekell J, Peters J, et al. Does participation in psychosocial treatment augment the benefit of clozapine? Department of Veterans Affairs Cooperative Study Group on Clozapine in Refractory Schizophrenia. Arch Gen Psychiatry. 1998;55(7):618-625.

23. Meltzer HY, McGurk SR. The effects of clozapine, risperidone, and olanzapine on cognitive function in schizophrenia. Schizophr Bull. 1999;25(2):233-255.

24. Glick ID. Understanding the results of CATIE in the context of the field. CNS Spectr. 2006;11(7 suppl 7):40-47.

25. Rosenheck RA. Effectiveness versus efficacy of second-generation antipsychotics: haloperidol without anticholinergics as a comparator. Psychiatr Serv. 2005;56:85-92.

26. Jibson MD, Tandon R. New atypical antipsychotic medications. J Psychiatr Res. 1998;32(3-4):215-228.

27. Wang PS, Scheeweiss S, Avorn J, et al. Risk of death in elderly users of conventional vs. atypical antipsychotic medications. N Engl J Med. 2005;353(22):2335-2341.

28. Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH. The disease burden associated with overweight and obesity. JAMA. 1999;282(16):1523-1529.

29. Kushner RF, Foster GD. Obesity and quality of life. Nutrition. 2000;16(10):947-952.

30. Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW Jr. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med. 1999;341(15):1097-1105.

31. Wetterling T, Mussigbrodt HE. Weight gain: side effect of atypical neuroleptics? J Clin Psychopharmacol. 1999;19(4):316-321.

32. Taylor DM, McAskill R. Atypical antipsychotics and weight gain—a systematic review. Acta Psychiatr Scand. 2000;101(6):416-432.

33. Sernyak MJ, Leslie DL, Alarcon RD, Losonczy MF, Rosenheck R. Association of diabetes mellitus with use of atypical neuroleptics in the treatment of schizophrenia. Am J Psychiatry. 2002;159(4):561-566.

34. Bustillo J, Lauriello J, Horan W, Keith S. The psychosocial treatment of schizophrenia: an update. Am J Psychiatry. 2001;158(2):163-175.

35. Wu EQ, Birnbaum HG, Shi L, et al. The economic burden of schizophrenia in the United States in 2002. J Clin Psychiatry. 2005;66(9):1122-1129.

36. Ritsner M, Gibel A, Perelroyzen G, Kurs R, Jabarin M, Ratner Y. Quality of life outcomes of risperidone, olanzapine, and typical antipsychotics among schizophrenia patients treated in routine clinical practice: a naturalistic comparative study. J Clin Psychopharmacol. 2004;24(6):582-591.

37. Lieberman JA. Comparative effectiveness of antipsychotic drugs. A commentary on: Cost Utility Of The Latest Antipsychotic Drugs In Schizophrenia Study (CUtLASS 1) and Clinical Antipsychotic Trials Of Intervention Effectiveness (CATIE). Arch Gen Psychiatry. 2006;63(10):1069-1072.

38. Rosenheck RA. Outcomes, costs and policy caution. A commentary on: the Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study (CUtLASS 1). Arch Gen Psychiatry. 2006;63(10):1074-1076.

39. Abulseoud O, Fayek M, Kingsbury SJ, Simpson GM. Patients’ preference for conventional antipsychotic medications. Psychiatr Serv. 2002;53(5):537-538,547.