The term atypical antipsychotics alludes to the existence of a class of drugs which are supposed to be at least as effective as the old, typical antipsychotics for antipsychotic effect but are less likely to cause severe side effects, especially fewer extrapyramidal symptoms (EPS), including tardive dyskinesia.(1–3) Atypical antipsychotics are also referred to as second-generation antipsychotics.
The term atypical antipsychotic met widespread acceptance in the 1990s when clozapine was reintroduced in Europe and North America for the treatment of “treatment-resistant schizophrenia,” that is, schizophrenia not responding to typical, first-generation antipsychotics.(4,5) Two decades earlier, clozapine had already been used as a first-line treatment for schizophrenia, but in 1975, it was withdrawn due to the occurrence of agranulocytosis. Clozapine’s mode of action is different from other antipsychotics in that its clinical potency does not correlate with its ability to block dopamine D2 receptors.(4,6) Also, the lower rate of EPS matched the idea of a different mode of action.
Subsequently, the “new” and “atypical” mode of action of clozapine was attributed to several new drugs.(1,2) At the moment, the following drugs are considered “atypical” antipsychotic agents: clozapine (marketed as Clozaril), quetiapine (Seroquel), risperidone (Risperdal), olanzapine (Zyprexa), ziprasidone (Geodon), aripiprazole (Abilify), melperone (Buronil), asenapine (Saphris), blonanserin (Lonasen), iloperidone (Fanapt), lurasidone (Latuda), and sertindole (Serdolect). However, there is no universally accepted definition of “atypical” and some of these drugs are clinically indistinguishable in their movement side effects from first-generation antipsychotics.
Meltzer, Matsubara, and Lee hypothesized that the favorable clinical characteristic of clozapine, that is, the low risk for EPS, is due to the relatively stronger 5-HT2A-receptor affinity compared with that for the dopamine D2 receptor.(7) This concept largely contributed to the development of the leading atypical antipsychotics such as risperidone, olanzapine, quetiapine, and ziprasidone.(6) Aripiprazole is different in this respect because it is a partial dopamine D2 receptor agonist (and thus a weak D2 receptor agonist) and also a 5-HT2A-antagonist.(8) Kapur and Remington challenged this hypothesis and suggested that the factor responsible for atypical antipsychotics being atypical is “fast dissociation” from the D2 receptor, leading to transient or easily displaceable occupancy.(9) Westerink proposed moderate dopamine D2 receptor occupancy as another key feature.(10) This might be the case for clozapine and quetiapine, which have low affinities for the D2 receptor.(6,11)
When analyzing or comparing studies of movement disorders associated with atypical antipsychotics, it is important to consider several methodological issues that potentially limit conclusions. For example, the incidence and severity of movement disorders can be affected by the following: drug dose, concomitant administration of antiparkinsonian drugs, the duration of treatment and observation, the extent of prestudy washout periods and carryover effects of prior drug treatment, and the susceptibility of the patient sample (e.g., the elderly, drug-naïve or treatment refractory patient).(12) Another important factor to consider is the number of dropouts, which, in general, is large in these kinds of studies.(13)
When assessing the effect of atypical antipsychotics in comparison with typical agents, it is important to consider studies using low- as well as high-potency drugs as comparators, because side effects of first-generation antipsychotics typically vary with the potency of the agent. Whether a drug is classified as high or low-potency is determined by its relative antipsychotic effectiveness compared to “chlorpromazine.”(14,15) Examples of high-potency drugs are fluphenazine and haloperidol. High-potency antipsychotics are associated with more frequent EPS and less histaminic effects (e.g., sedation), alpha-adrenergic effects (e.g., orthostatic hypotension), and anticholinergic effects.(16) In contrast, low-potency agents have fewer EPS, a high risk of sedation, a high risk of orthostatic hypotension and tachycardia, and a high risk of anticholinergic and antiadrenergic effects.(16)
Early randomized trials suggested that atypical antipsychotics are superior to first-generation antipsychotics in the treatment of schizophrenia.(17–20) More recent studies, however, largely conducted by independent investigators, have found less substantial differences in EPS between these two classes of drugs.(21,22) The differences between the earlier studies and more recent studies may be related to the earlier studies commonly utilizing the high-potency haloperidol, often at higher doses, as the comparator.(22,23) For example, in clinical trials risperidone resulted in a significantly lower incidence of acute EPS, greater reductions in EPS rating scores, fewer dropouts, and less frequent use of anticholinergic drugs compared with haloperidol.(24–27) In a small study of first-episode schizophrenics, risperidone-treated patients experienced one-fourth the incidence of parkinsonism and concurrent anticholinergic use compared with patients receiving typical antipsychotics.(28) However, postmarketing studies and closer inspection of trial data show that the occurrence of EPS is dose dependent. Advantages of risperidone disappear if doses above 6 mg per day are used or if lower doses of haloperidol or less potent drugs are used in comparison.(22,29) The CATIE trial found no differences in EPS between typical and atypical antipsychotics.(21)
Over time, it has become clear that the atypical antipsychotics are not a homogeneous class and differ in many properties. With the exception of clozapine, olanzapine, and risperidone, atypical antipsychotics are not better than low-potency first-generation antipsychotic drugs.(30) Nevertheless, the distinguishing features that are attributed to clozapine seem to withstand the current debate about whether atypical antipsychotic drugs are better than first-generation antipsychotics.
In concordance with the above, the risk for late or tardive dyskinesias appears to be lower with atypical antipsychotics than with the older antipsychotics. However, risks of tardive syndromes with atypical antipsychotics have been reduced much less than expected.(31) In a review of 12 trials involving 28,051 patients (followed for 463,925 person-years) the annualized tardive dyskinesia incidence was 3.9% for atypical antipsychotics and 5.5% for typical antipsychotics.(32) Stratified by age, annual tardive dyskinesia incidence rates were 3.0% with atypical antipsychotics versus 7.7% with typical antipsychotics in adults, and 5.2% for both atypical and typical antipsychotics in the elderly (based almost exclusively on one retrospective cohort study). A cautious interpretation of these results is warranted because the review was based on a mix of controlled and uncontrolled studies and blinded and open-label studies, as well as historical and limited direct comparisons with typical antipsychotics (the latter unfortunately exclusively with medium to high doses of haloperidol).(32,33)
Optimistic expectations for the efficacy and neurological safety of atypical antipsychotics have encouraged their wide use in many conditions, sometimes off-label or in combinations, increasing the chance of a higher prevalence of tardive syndromes, especially in older ages.(31) The risk for tardive syndromes is low with clozapine.(34) Data are insufficient to support or refute use of atypical antipsychotics to improve tardive syndromes.(35)
Although neuroleptic malignant syndrome (NMS) is historically associated with the typical antipsychotic drugs, it is also a potential adverse effect of atypical antipsychotics. Troller et al. reviewed published cases of NMS with the following atypical antipsychotics: clozapine, quetiapine, risperidone, olanzapine, ziprasidone, and aripiprazole.(36) The majority of cases of NMS induced by atypical antipsychotics present with the typical NMS features. There appears to be one major exception, clozapine-induced NMS is significantly less likely than NMS induced by risperidone, olanzapine, and typical antipsychotics to manifest with rigidity.(36) Incidence estimates for NMS vary widely, and it remains uncertain whether atypical drugs are less likely to cause NMS than the typical antipsychotics. Caroff and Mann have estimated the incidence of NMS with typical antipsychotics to be approximately 0.2%.(37) This low incidence makes it unlikely that comparative studies will be adequately powered to detect real differences in the incidence of NMS induced by atypical versus typical antipsychotics.
Clozapine is a dibenzodiazepine derivate that was developed in the 1950s. Its clinical use had been delayed because of an associated risk of agranulocytosis.(12) Following the publication by Kane et al. in 1988 of a multicenter clinical trial that assessed clozapine’s efficacy compared to chlorpromazine in the treatment of patients who were refractory to typical neuroleptics, clozapine received renewed attention.(38) In this study, 30% of the clozapine-treated patients were categorized as responders compared with 4% of chlorpromazine-treated patients. The authors also found a reduction of EPS, which was significantly greater in the clozapine group compared to the patients receiving chlorpromazine combined with benztropine mesylate.(38) Although not a study of new onset or actual incidence of EPS, this report supported a more favorable reduction in the prevalence of preexisting EPS with clozapine.(12)
In a Cochrane review of randomized controlled trials (RCT) comparing clozapine with typical antipsychotics for schizophrenia, the risk ratio for the development of motor adverse effects with clozapine was 0.57 (95%CI 0.50 to 0.65) when compared to typical antipsychotics.(39) For children and adolescents this ratio was 0.77 (95%CI 0.67 to 0.90) and for the elderly 0.75 (95%CI 0.22 to 2.60). The study included 52 trials (4,746 participants). Forty-four of the included studies were less than 13 weeks in duration, and, overall, trials were at a significant risk of bias.(39) In a Cochrane review of clozapine versus the other atypical antipsychotics, clozapine tended to cause less EPS than the other atypical antipsychotics, but this was only significant compared to risperidone for parkinsonism with a risk ratio of 0.39 (95%CI 0.22 to 0.68) and to zotepine with a risk ratio of 0.05 (95%CI 0.00 to 0.86).(40)
A very common nonmotor symptom of Parkinson’s disease is visual hallucinations, with prevalence rates ranging from 22%–38% and a lifetime prevalence of up to 70%.(41) Less common, but more troubling, are delusions, which are usually paranoid in nature, and affect about 5% of drug-treated PD patients. Unfortunately, all of the drugs available for treating motor symptoms in Parkinson’s disease worsen the psychotic symptoms. Levodopa, which is the most potent and least expensive antiparkinson drug, is least likely to contribute to the hallucinations.(41)
When visual hallucinations become troublesome—due to loss of insight—antipsychotic drugs are prescribed. In spite of the established effectiveness of clozapine, proven by two multicenter, double blind, placebo controlled studies for the treatment of Parkinson’s disease psychosis, several drawbacks limit the earlier use when hallucinations are still minor.
In the 4-week double blind RCT, conducted by the Parkinson Study Group, 60 patients with Parkinson’s disease and psychosis were randomized to low doses of clozapine or placebo.(42) The patients in the clozapine group had significantly more improvement than those in the placebo group on all three of the measures used to determine the severity of psychosis. Clozapine treatment had no deleterious effect on the severity of parkinsonism and even improved tremor. In a subsequent 12-week, open-label extension of the study there was an unexpectedly high death rate.(43) The authors state that the death of six patients was a result of the nature of the advanced disease and not of clozapine. In 2004, Pollak et al. conducted a 4-week, double blind RCT of clozapine and placebo, followed by a 12-week clozapine open phase.(44) This was then followed by a 1-month washout period. At the end of the 12-week open period, 25 of the initial 55 patients had completely recovered from delusions and hallucinations and of these, 19 experienced a relapse within 1 month of the clozapine washout period.
Cumbersome monitoring, consisting of frequent white blood cell counts, is required when using clozapine because of the 1%–5% incidence of agranulocytosis in the elderly.(40,45–47) This hematological problem is not dose related and therefore frequent monitoring needs to be done, even with doses as low as 6.25 mg per day (N.B.—usually clozapine doses in schizophrenia are 300–900 mg per day). Weight gain and the metabolic syndrome are not a problem in Parkinson’s disease patients, probably because of the low doses.(40,44) Other side effects such as sedation and orthostatic hypotension are frequently observed and may be a dose-limiting factor. The sedating effect of clozapine may be used in an advantageous way because many psychotic patients with Parkinson’s disease are awake at night when their symptoms worsen. As a result, clozapine is typically administered at bedtime. This is sometimes supplemented with a lower dose in the morning. A normal starting dose is 6.25 mg at bedtime, which then may be increased every one to two weeks with 6.25 mg or 12.5 mg per day. On average, doses are between 12.5 mg and 50 mg per day.
Quetiapine, a dibenzothiazepine compound, was developed specifically to have a profile similar to clozapine with respect to EPS, but without the risk of hematological problems. Like clozapine, it has a greater affinity for 5-HT2A than dopamine D2 receptors.(6,12)
In patients with schizophrenia, quetiapine causes fewer EPS compared with typical antipsychotics (risk ratio 0.17; 95%CI 0.09 to 0.32), including less akathisia, parkinsonism, dystonia, and tremor.(48) When compared to clozapine, there is no difference regarding the frequency of akathisia, rigidity, tremor, or use of antiparkinson medication.(49) In comparison to olanzapine, fewer patients on quetiapine need to take antiparkinson medication at least once (risk ratio 0.51, 95%CI 0.32 to 0.81), but apart from this, there are no significant differences in EPS. Quetiapine produces fewer EPS than risperidone when assessed with the Simpson-Angus Scale. There are no significant differences regarding dyskinesia and akathisia. When compared to ziprasidone, significantly fewer people in the quetiapine group use antiparkinson medication at least once. The frequency of akathisia or “any EPS” are the same. Studies comparing quetiapine with aripiprazole do not show a difference regarding frequency of EPS, but there are only a few studies, and those comprise small sample sizes.(49)
The first successful report on quetiapine use in parkinsonian patients was published in 1996.(50) More open studies of patients with Parkinson’s disease followed and these suggested an advantage of quetiapine in this population; that is, improvement in psychosis was seen in more than 80% of patients, and motor worsening in only 13%.(51,52) Consistently reported side effects of quetiapine include sedation and hypotension.(53)
In 2005, Ondo et al. published the results of a double blind RCT of quetiapine for hallucinations in Parkinson’s disease.(54) Follow-up was 12 weeks. Among the 31 patients originally included in the trial, there were six dropouts. Compared to placebo, none of the hallucinations, psychosis, or motor impairment assessments changed significantly on quetiapine. Rabey et al. conducted a double blind 12-week RCT investigating a total of 58 Parkinson’s disease patients of whom 29 were demented.(55) The study was characterized by a high dropout rate of 45% (n = 26) primarily due to lack of efficacy. Compared to placebo no outcome variable changed significantly with quetiapine. This was also true when considering only the demented or the nondemented patients. Shotbolt et al. published the results of their RCT of quetiapine in 24 subjects in 2009.(56) Thirteen patients completed 6 weeks of the study and eight completed the 12-week double blind phase. There was no significant difference in time-to-dropout between patients receiving quetiapine or placebo. No significant changes were found for any of the endpoints in either group. Fernandez et al. (2009) conducted a 1-month double blind RCT of quetiapine for the treatment of visual hallucinations, using changes in REM sleep architecture (assessed with polysomnography) as the primary endpoint.(57) Eleven of the 16 patients randomized completed the study. There was no statistically significant difference in change of REM sleep duration in either arm.
Two studies compared quetiapine with clozapine for the treatment of psychosis in patients with Parkinson’s disease. Morgante et al. (2004) conducted a 12-week rater-blinded RCT to compare quetiapine and clozapine in 20 patients.(58) No significant differences were found in the psychosis and motor scores between the two groups. In 2006, Merims et al. published the results of a 22-week RCT.(59) Only seven of the 14 patients randomized to receive clozapine and nine of the 13 randomized to quetiapine completed the study. Compared to baseline, clozapine, but not quetiapine, significantly improved the frequency scores of the categories “hallucinations” and “delusions” of the neuropsychiatric inventory (NPI). The severity scores of hallucinations and delusions in either group did not change significantly. There was no worsening in parkinsonian symptoms.
For some time, quetiapine had been the treatment of first choice for Parkinson’s disease psychosis because of ease of use, lack of compulsory blood monitoring, and consistently encouraging results in open-label studies.(51) But because of the not so positive efficacy data in recent RCTs and several methodological concerns of these studies (e.g., small sample size, low quality rating), there is insufficient evidence to strongly support the use of quetiapine for the treatment of psychosis in Parkinson’s disease. Importantly, there are no new safety concerns identified in the new studies.(53)
For psychosis in Parkinson’s disease, quetiapine can be started at, for example, 12.5 mg per day (bedtime) and then increased with, for example, 25 mg per day each week until it causes either symptomatic effects or adverse effects. In general, dosages range from 25 to 200 mg per day.
Risperidone, a benzisoxazole, was introduced in 1994. As with the other “atypicals” it was initially hoped that risperidone would have the same favorable EPS profile as clozapine. And although risperidone is an effective and widely used antipsychotic agent, it has not lived up to the expectation that it would provide an alternative to clozapine.
Initial studies with risperidone for schizophrenia were quite promising, with lower rates of EPS, decreased dropout rates, and a reduction in the concomitant use of anticholinergics compared with haloperidol treatment.(26,27,60–64) These benefits were soon found to be highly dose dependent; that is, the advantages of risperidone were greatly diminished at doses greater than 6 mg per day.(26,27,60–64) This dose dependency was also seen in drug-naïve patients, and not only with respect to the occurrence of EPS, but also regarding the use of anticholinergic medications. When lower doses of haloperidol or low-potency typical antipsychotics were compared to risperidone, the EPS liability of risperidone was similar.(26,27,60–65)
There are not enough data available to allow for a reliable comparison of risperidone and clozapine regarding EPS.(66) When risperidone is compared to olanzapine, those taking olanzapine have slightly fewer EPS in the medium term (risk ratio 1.67, 95%CI 1.14 to 2.46), have fewer new episodes of parkinsonism (risk ratio 1.73, 95%CI 1.07 to 2.81), and need less antiparkinsonian medication (risk ratio 1.67, 95%CI 1.15 to 2.41).(66) In comparison to amisulpride, there are no significant differences with risperidone in terms of either the number of patients who reported EPS or received antiparkinsonian medication.(66)
Risperidone reduces hallucinations and psychosis in Parkinson’s disease. However, several open-label series have demonstrated consistently that the drug worsens motor features of Parkinson’s disease.(67,68) One double blind RCT compared clozapine and risperidone in 10 patients with Parkinson’s disease.(69) The study demonstrated similar efficacy in both treatment groups for psychiatric symptoms, but there was nonsignificant worsening of motor symptoms in the risperidone treatment group compared to improvement of motor symptoms in the clozapine group. In conclusion, risperidone is not useful for the treatment of psychosis in Parkinson’s disease.
Olanzapine, a thienobenzodiazepine derivative, is the next atypical antipsychotic agent introduced after risperidone. Patients with schizophrenia taking olanzapine experience fewer EPS than those given typical antipsychotics. Olanzapine may be associated with slightly more extrapyramidal side effects than quetiapine, but less than risperidone and ziprasidone.(71)
In 2002, Breier et al. reported two double blind, placebo controlled RCTs that examined the efficacy of low-dose olanzapine for psychosis in patients with Parkinson’s disease.(72) Both studies, one in the United States (n = 83) and the other in Europe (n = 77), were reported in one publication. Although no significant treatment group differences regarding psychosis were observed, motor function worsened significantly in patients on olanzapine compared to placebo in both trials. In the U.S. study, patients in the olanzapine group showed significantly higher rates of EPS (olanzapine, 24.4%; placebo, 2.4%; P = .003), hallucinations (olanzapine, 24.4%; placebo, 4.8%; P = .013), and increased salivation (olanzapine, 22.0%; placebo, 4.8%; P = .026).(72)
In the same year, Ondo et al. published a smaller double blind, placebo controlled RCT.(73) Thirty patients with Parkinson’s disease and hallucinations underwent nine weeks of treatment with olanzapine or placebo (2:1 ratio). This study also failed to detect significant differences between olanzapine and placebo in any of the psychosis measures. Again, there was significant worsening of motor symptoms in the olanzapine group, mainly through the worsening of gait and bradykinesia.