Adverse Effects of Atypical Antipsychotics
Antipsychotic drugs can be of great benefit in a range of psychiatric disorders, including schizophrenia and bipolar disorder, but all are associated with a wide range of potential adverse effects. These can impair quality of life, cause stigma, lead to poor adherence with medication, cause physical morbidity and, in extreme cases, be fatal. A comprehensive overview of tolerability requires a review of all available data, including randomised controlled trials (RCTs), observational studies and postmarketing surveillance studies. Assessing the relative tolerability of atypical antipsychotics is hampered by the paucity of RCTs that compare these drugs head-to-head, and limited and inconsistent reporting of adverse effect data that makes cross-study comparisons difficult.
Despite methodological problems in assessment and interpretation of tolerability data, important differences exist between the atypical antipsychotics in the relative risk of acute extrapyramidal symptoms (highest risk: higher doses of risperidone), hyperglycaemia and dyslipidaemia (highest risk: clozapine and olanzapine), hyperprolactinaemia (highest risk: amisulpride and risperidone), prolongation of heart rate-corrected QT interval (QTc) [highest risk: ziprasidone and sertindole] and weight gain (highest risk: clozapine and olanzapine). Sedation, antimuscarinic symptoms, postural hypotension, agranulocytosis and seizures are more common with clozapine than with other atypical antipsychotics. The variation in their tolerability suggests that it is misleading to regard the atypical antipsychotics as a uniform drug class, and also means that the term ‘atypical antipsychotic’ has only limited usefulness. Differences between the atypical agents in terms of efficacy and pharmacodynamic profiles also support this view.
As tolerability differs between specific conventional and atypical drugs, we conclude that broad statements comparing the relative risk of specific adverse effects between ‘atypical’ and ‘conventional’ antipsychotics are largely meaningless; rather, comparisons should be made between specific atypical and specific conventional drugs. Adverse effects are usually dose dependent and can be influenced by patient characteristics, including age and gender. These confounding factors should be considered in clinical practice and in the interpretation of research data. Selection of an antipsychotic should be on an individual patient basis. Patients should be involved in prescribing decisions and this should involve discussion about adverse effects.
The effectiveness of atypical antipsychotics in acute schizophrenia and in preventing relapse is clearly demonstrated in meta-analyses. Most treatment guidelines recommend atypical antipsychotics in preference to conventional antipsychotics1 in those patients with newly diagnosed schizophrenia. A strong evidence base supports the use of atypical antipsychotics in mania, with several agents now licensed for this indication. Emerging data support the use of certain atypical drugs in other conditions, including bipolar depression and the prevention of relapse in bipolar disorder. Atypical antipsychotics, like all drugs, are associated with adverse effects; however, until better tolerated or more effective drugs are developed, they are likely to remain a key treatment option in schizophrenia and other major psychiatric disorders. When considering the adverse effects of atypical antipsychotics it is important to consider the severity of the illnesses in which they are indicated and their benefits when prescribed appropriately. Schizophrenia and bipolar disorder both rank among the leading causes of disability in the world.
This review outlines the clinical relevance of adverse effects, the range of atypical drugs currently available in the UK, sources of adverse effect data and methodological limitations of the current evidence base. Specific adverse effects are considered, with the emphasis on the relative risk with different atypical antipsychotics. Where relevant, comparisons are made with the risk seen with specific conventional antipsychotics. The differential diagnosis and treatment of adverse effects are beyond the scope of this review, but are dealt with in a recent textbook to which readers are referred.
1. Clinical Consequences and Methodological Issues
1.1 Clinical Consequences of Adverse Effects
Adverse effects are important for the following reasons:
They can cause distress to patients and impair quality of life.
They can cause poor adherence with medication, which may contribute to a poor response during short-term treatment or a relapse during maintenance treatment.
They can be stigmatising; for example, weight gain and abnormal involuntary movements are easily observable by others and may mark the patient as ‘different’.
Adverse effects may lead to physical morbidity and increased mortality; for example, weight gain is a risk factor for ischaemic heart disease, prolongation of heart rate-corrected QT interval (QTc) is associated with ventricular arrhythmias and sudden unexpected death, and postural hypotension can cause falls and hip fractures.
Knowledge of how the prevalence and severity of adverse effects varies between specific antipsychotics allows the clinician to minimise their occurrence. Familiarity with adverse effects also facilitates their recognition and management.
1.2 Tolerability and Patient Characteristics
Patient characteristics can influence the relative risk of adverse drug effects. For example, the elderly are more prone to develop adverse effects for several reasons. First, compared with younger adults, the elderly tend to have a lower body mass, a higher proportion of body fat to body water, and decreased renal and hepatic function. All of these factors indicate that, at any given dose, an elderly person is more likely to have a higher plasma drug concentration than a younger person. Second, the elderly are more sensitive to many drug effects due to increased receptor sensitivity and less responsive regulation of physiological parameters such as temperature and blood pressure. Finally, elderly people are more likely to take multiple drugs, thereby increasing the risk of adverse effects secondary to drug interactions. Examples of antipsychotic-induced adverse effects that are more common in the elderly include tardive dyskinesia, QTc prolongation, clozapine-induced agranulocytosis and postural hypotension.
Other patient characteristics that influence the risk of adverse effects are gender and co-morbid physical illness. Women are more prone than men to develop hyperprolactinaemia and QTc prolongation when treated with antipsychotics, and more prone to develop agranulocytosis when treated with clozapine. Those with a history of epilepsy are more likely to experience seizures with antipsychotics, particularly clozapine.
In clinical practice, patient characteristics may influence the choice of antipsychotic drug, the starting dose and the rate at which dose is increased. At an academic level it is important to ensure that trial populations are comparable if one attempts to make cross-study comparisons of tolerability.
1.3 Range of Atypical Antipsychotics
Atypical antipsychotics are generally defined as drugs that cause minimal or no extrapyramidal symptoms (EPS) at therapeutic doses. The distinction from conventional antipsychotics is not absolute. For example, thioridazine, a so-called conventional antipsychotic, has little propensity to cause EPS. Several pharmacological mechanisms may account for ‘atypical’ characteristics, including the degree of antagonism at dopamine D2 receptors, the speed of dissociation from the D2 receptor, the antimuscarinic potential of the drug and the degree of serotonin 5-HT2A receptor antagonism. Given that different mechanisms can underlie atypicality, it is not surprising that atypical antipsychotics differ in terms of chemical structure, pharmacological actions and adverse effect profiles.
The number of licensed atypical drugs has increased over the last 10 years and further agents are in development. Table I lists the atypical antipsychotics currently licensed in the UK and their UK indications. Some of these drugs are not licensed in other countries but other agents are available, e.g. amisulpride is licensed in the UK but not in the US and, conversely, ziprasidone is licensed in the US but not in the UK.
1.4 Sources of Data
Data on the adverse effects of atypical antipsychotics are available from a range of sources, including randomised controlled trials (RCTs), observational studies and postmarketing surveillance. Each has its own strengths and weaknesses and the best overview of tolerability comes when all data sources are considered.
Strengths of RCTs include (i) randomisation, which reduces the risk of bias in baseline characteristics, making it more probable that differences in outcome reflect differences between treatments; and (ii) prospective design that allows accurate assessment of outcome, in this case adverse effects. RCTs may record spontaneously reported adverse events and/or employ specific scales to rate adverse effects. Rating scales improve the reliability and validity of the assessment but the interpretation of the resulting data is still subject to methodological limitations, as discussed in section 1.5.
A weakness of RCTs is the generalisability of the data. Extrapolating from RCTs to clinical practice is limited by the short duration of most antipsychotic trials (≤6 weeks), which means they cannot provide data on long-term adverse effects; for example, tardive dyskinesia or long-term weight gain. Most RCTs employ multiple exclusion criteria, with the result that participants are not representative of most ‘real world’ patients.
The CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) study is more representative of clinical practice than many RCTs because of its relatively long duration and minimal exclusion criteria. CATIE has the additional benefits of being independent of the pharmaceutical industry and a large sample size. In phase I of the CATIE study nearly 1500 patients with schizophrenia were randomised to one of four atypical antipsychotics (olanzapine, quetiapine, risperidone and ziprasidone) and perphenazine, a conventional antipsychotic, and followed up for 18 months. The aim of this phase of the study was to assess overall antipsychotic effectiveness, and the primary outcome measure was discontinuation of treatment for any reason.
Observational studies have the advantage of assessing ‘real world’ patients and often have longer follow-up periods than RCTs. However, the assessment of adverse effects is often less rigorous than in RCTs and the lack of randomisation means one cannot exclude a prescribing bias.
Postmarketing surveillance is an important source of tolerability and safety data, and includes prescription event monitoring and adverse drug reaction (ADR) reports. Various national and international regulatory bodies provide systems for postmarketing surveillance, an example being the UK ‘yellow card’ system for reporting ADRs. A major problem with postmarketing surveillance data is that it is difficult to attribute causality. ADR reporting systems have the additional problems of widespread under-reporting and an inability to reliably assess the incidence of adverse effects.
1.5 Methodological Limitations
A range of methodological issues hamper interpretation of the literature on antipsychotic tolerability. Few studies compare atypical antipsychotics head-to-head. Many papers that report RCTs provide little tolerability data and instead concentrate on efficacy. Where adverse effect data are provided, they are often limited and presented inconsistently, making cross-study comparisons difficult. Outcome measures often lack clinical usefulness. For example, some trials report the mean change in plasma glucose and serum prolactin levels from baseline to endpoint, measures that have little relevance to patients or clinicians. In contrast, reporting the number of patients who moved from having ‘normal range’ plasma/serum levels at baseline to ‘above the upper limit of normal’ levels at endpoint has far more clinical utility.
Most trials of atypical antipsychotics evaluate patients with chronic psychosis who discontinued an antipsychotic before starting the trial, making drug carry-over effects inevitable. For example, the weight gain potential of an antipsychotic is underestimated in this design as patients are likely to have gained weight during previous antipsychotic treatment. Assessing patients with first-episode psychosis, who have not received prior antipsychotic treatment overcomes this problem, but few RCTs have been conducted in this group. Industry-sponsored trials are more likely to report results that favour the sponsor’s compound than are independent studies. This may reflect publication bias or bias in the design of the trial. Finally, for some adverse effects, for example dystonia, there are no recognised rating scales. A more detailed discussion of these methodological issues and possible solutions is provided in a review by Hamer and Haddad.
2. Extrapyramidal Symptoms
2.1 Parkinsonism, Akathisia and Acute Dystonia
Antipsychotic drugs cause four main extrapyramidal syndromes – parkinsonism, akathisia, acute dystonia and tardive dyskinesia. The first three syndromes usually appear within the initial weeks of treatment, are considered in this section and, for convenience, will be referred to as EPS in further discussion. Tardive dyskinesia is usually associated with longer term treatment and is dealt with separately in section 2.2.
EPS occur in up to 75% of patients treated with conventional antipsychotics. The associated subjective discomfort is cited as a major cause of poor compliance. Although parkinsonism and dystonia can be treated with anticholinergic drugs, and akathisia can be treated with benzodiazepines and propranolol, these treatments have their own adverse effects and contraindications. Anticholinergic drugs have a misuse potential and can cause various adverse effects, including delirium. Benzodiazepines are associated with tolerance and dependence, and propranolol is contraindicated in those with asthma and peripheral vascular disease.
A reduced risk of EPS, compared with conventional antipsychotics, has been demonstrated in RCTs for all the atypical antipsychotics listed in table I.
This advantage is also evident in meta-analyses.[26,27] For example in one meta-analysis between 11% and 57% of patients prescribed an atypical antipsychotic required antiparkinsonian medication compared with between 36% and 81% of patients prescribed conventional antipsychotics. An important proviso is that in most RCTs the conventional comparator is haloperidol, a drug associated with a high rate of EPS as a result of its high potency for blocking D2 receptors. Such trials can be regarded as having a design bias that favours the atypical antipsychotic.
Studies that compare atypical antipsychotics with conventional antipsychotics other than haloperidol, reveal a less marked difference in EPS propensity. These studies include CATIE, CUtLASS-1 (Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study) and a meta-analysis by Leucht et al. In phase I of the previously described CATIE study, the incidence of EPS did not differ significantly between perphenazine and the atypical drugs over the 18-month follow-up; however, more patients discontinued perphenazine (8%) than atypical antipsychotics (2–4%) as a result of EPS, implying a greater severity, if not overall incidence, of EPS with perphenazine. In the CATIE study, patients with tardive dyskinesia at baseline were excluded from randomisation to perphenazine, introducing a possible bias that favours perphenazine. In the CUtLASS-1 study, patients were randomised to a conventional or atypical antipsychotic, with the choice of individual drug being made by the managing clinician. The most frequently used conventional drug was sulpiride, a medium potency drug. The incidence of EPS did not differ between the atypical and conventional cohorts during the 12-month follow-up. The meta-analysis by Leucht et al. compared atypical drugs with low-potency conventional drugs (equivalent or less potent than chlorpromazine) and found that mean dosages of chlorpromazine <600 mg/day or its equivalent had no higher risk of EPS than atypical drugs, other than clozapine. However, conventional drugs at a dosage exceeding 600 mg/day of chlorpromazine equivalents were associated with a higher incidence of EPS than atypical drugs. Nevertheless, owing to the relatively short duration of the trials, CATIE, CUtLASS and the Leucht meta-analysis are all unable to assess the risk of tardive dyskinesia during long-term treatment with atypical antipsychotics.
The EPS advantage of atypical antipsychotics led the UK National Institute of Clinical Excellence (NICE) Schizophrenia Guideline to recommend atypical antipsychotics in preference to conventional antipsychotics as first-line agents in patients with new-onset schizophrenia. Worldwide, many other guidelines make this recommendation. NICE also recommended atypical antipsychotics for patients treated with conventional antipsychotics who had unacceptable adverse effects, irrespective of the degree of symptom control.
Atypical drugs differ in their relative risk of EPS. A meta-analysis concluded that risperidone, within the optimum daily dosage range, was associated with less frequent EPS and less frequent use of anticholinergic medication than conventional antipsychotics, mainly haloperidol. However, at higher dosages of risperidone (>6 mg/day) this difference disappears.[32,33] Amisulpride also causes dose-dependent EPS. A Cochrane review reported that the use of medication to alleviate EPS was significantly more common with risperidone than olanzapine (25% vs 18%). Clozapine and quetiapine demonstrate placebo-level incidence of EPS across their dosage range.[36,37] Aripiprazole and zotepine both cause low rates of EPS, although the number of published studies for each drug is small. During an 8-week trial, no significant difference in parkinsonian symptoms with zotepine compared with placebo or chlorpromazine was shown.
In summary, meta-analyses show an EPS advantage for the atypical antipsychotics compared with conventional antipsychotics, although in most cases the comparator is haloperidol.[26,27] Within the atypical class, the relative risk of EPS varies, particularly at higher doses. Amisulpride and risperidone show evidence of a dose-related increase in EPS, but clozapine and quetiapine demonstrate placebo-level incidence across their dosage range. Since clozapine requires ongoing haematological monitoring, some experts have suggested that quetiapine ought to be the drug of choice for the treatment of psychosis in EPS-vulnerable patients.
2.2 Tardive Dyskinesia
Tardive dyskinesia is characterised by various involuntary movements, including myoclonic jerks, tics, chorea and dystonia. Most commonly it affects the oro-facial muscles, but virtually any part of the body can be affected and it may persist despite drug withdrawal. With conventional drugs, the probability of developing tardive dyskinesia is approximately 5% per year over the first decade of treatment.[41,42] Beyond this time the incidence continues to rise, although less steeply. Older patients are at increased risk of developing this disorder. A naturalistic study of patients aged >55 years (mean age = 77 years) found that after 43 weeks of antipsychotic treatment the incidence of tardive dyskinesia was 31%, 6-fold higher than that reported in studies of younger patients. Patients with affective disorders appear at higher risk of developing tardive dyskinesia than those with schizophrenia.
Acute EPS is predictive of future development of tardive dyskinesia during treatment with conventional antipsychotics.[44,45] This suggests that the atypical antipsychotics, which are associated with a lower incidence of acute EPS than haloperidol, may be associated with a lower incidence of tardive dyskinesia. However, definitive data regarding the risk of tardive dyskinesia with atypical antipsychotics requires long-term studies, of which there are relatively few, to be conducted. Only two randomised, double-blind studies have assessed tardive dyskinesia beyond 1 year with an atypical drug versus a conventional comparator, in both cases haloperidol.[46,47] In both studies the median exposure time to the atypical antipsychotic was <1 year (260 days for olanzapine, 364 days for risperidone), indicating that most patients did not complete 1 year of treatment. In both studies, there was a lower risk of tardive dyskinesia for the atypical drug compared with haloperidol.[46,47] Many patients entering atypical antipsychotic trials have had prior treatment with conventional antipsychotics. As cumulative antipsychotic exposure is associated with the risk of developing tardive dyskinesia, such data may over-estimate the risk with atypical agents. Definitive data on the risk of tardive dyskinesia with atypical agents requires long-term studies of patients without prior conventional antipsychotic treatment, and at present such data are not available.
Correll et al. conducted a meta-analysis that included all open or controlled studies of at least 20 patients that lasted 1 year or longer, and reported on new cases of tardive dyskinesia. In 11 trials, a total of 2769 patients received treatment with an atypical antipsychotic (risperidone, olanzapine, quetiapine, amisulpride or ziprasidone). Four studies had a comparator — haloperidol in three, placebo in one. The weighted median duration of antipsychotic exposure was 306 days, indicating that most patients did not complete 1 year of treatment. In young and middle-aged adults (i.e. <54 years of age) the weighted mean annual incidence of tardive dyskinesia for the atypical antipsychotics was 0.8% (range 0.0–1.5%) compared with 5.4% (range 4.1–7.4%) in those treated with haloperidol. In older patients (>54 years of age) treated with atypical antipsychotics the incidence was 5.3% (range 0.0–13.4%), which, although higher than the risk in younger adults, remains approximately 5-fold less than that reported with conventional antipsychotics in this age group. This meta-analysis indicates that atypical drugs have a reduced risk of tardive dyskinesia compared with haloperidol, and that this benefit extends to elderly patients. An important proviso is that the mean haloperidol dosages in the three comparator studies in the meta-analysis were relatively high (>10 mg/day).
Dolder and Jeste assessed the incidence of tardive dyskinesia in a group of highly vulnerable patients, i.e. age ≥45 years with borderline tardive dyskinesia at baseline. After 6 months of treatment those treated with conventional antipsychotics were approximately twice as likely to develop tardive dyskinesia as those treated with an atypical antipsychotic (p < 0.001).
Several studies report improvement in tardive dyskinesia after patients are switched from a conventional to an atypical antipsychotic, including clozapine,[51, 52, 53] olanzapine, quetiapine and risperidone. Bai et al. used a randomised design to compare risperidone to placebo in the treatment of patients with schizophrenia and tardive dyskinesia. After 12 weeks the mean Abnormal Involuntary Movement Scale (AIMS) total score showed a significantly greater decrease in the risperidone group than in the placebo group. However, the relatively short duration means one cannot exclude the possibility of a withdrawal effect exacerbating tardive dyskinesia in the placebo arm, thereby inflating the difference between the placebo and risperidone arms.
Case reports have described tardive dyskinesia commencing during treatment with atypical antipsychotics, including clozapine, quetiapine, risperidone and olanzapine. Caution is needed in interpreting such reports for two reasons. First, most patients had received prior treatment with conventional antipsychotics, which will confer an increased risk for the future development of tardive dyskinesia. Second, dyskinesias can occur independent of antipsychotic use. As reviewed by Casey, spontaneous dyskinesias in patients with psychosis were recognised by both Bleuler and Kraepelin in the pre-antipsychotic era, and recent studies report a prevalence of 1–5% among never-medicated psychotic patients.[62,63] Jeste et al. reported a significant relationship between dose of risperidone and risk of tardive dyskinesia (p = 0.02), but there are no data on dose relationship for other atypical agents.
With regard to head-to-head data, an 18-week double-blind trial of patients with treatment-resistant schizophrenia reported a higher incidence of tardive dyskinesia with olanzapine than with clozapine (2.2% vs 0%; p < 0.03). However, a confounder is that by being treatment resistant, most patients had received conventional antipsychotics at high dosages and for long periods. The Correll et al. meta-analysis did not allow a comparison of the risk of tardive dyskinesia between different atypical drugs; however, of the 11 studies in the meta-analysis, the highest incidence of tardive dyskinesia (13.4% at 12 months) was seen in a risperidone trial in elderly patients.
In summary, a meta-analysis has shown that atypical drugs are associated with an approximately 5-fold lower risk of tardive dyskinesia than haloperidol during the first year of treatment. There is minimal head-to-head data comparing the risk of tardive dyskinesia between atypical drugs; however, based on existing data, and extrapolating from acute EPS data for the atypical drugs, it is reasonable to assume that the risk of tardive dyskinesia, across the full dosage range for each drug, is lowest with clozapine and quetiapine.
Antipsychotics cause hyperprolactinaemia by blocking D2 receptors on pituitary lactotroph cells, thereby removing the tonic inhibition on prolactin release provided by dopamine secreted by the hypothalamus. All conventional antipsychotics cause hyperprolactinaemia but the atypical drugs differ in their propensity to cause this effect. Aripiprazole causes a fall in plasma prolactin levels consistent with it being a D2 partial agonist. Clozapine and quetiapine do not elevate plasma prolactin levels across their full dosage range. Olanzapine and ziprasidone are regarded as prolactin sparing, although at higher dosages hyperprolactinaemia can occur. Data for zotepine are limited but plasma prolactin elevation is recognised. Risperidone and amisulpride cause a marked and sustained increase in serum prolactin levels in a sizeable proportion of patients.
Plasma prolactin elevation is more likely at higher antipsychotic dosages. However, individuals vary greatly in their susceptibility; some develop hyperprolactinaemia on low dosages of an antipsychotic whereas others maintain normal serum prolactin levels despite treatment with high dosages of the same drug. At any given antipsychotic dose, women are more likely to develop hyperprolactinaemia than men. Preliminary evidence, reviewed by Haddad and Wieck, suggests that postnatal women and adolescents are more vulnerable to develop hyperprolactinaemia.
In adults, hyperprolactinaemia tends to persist unless the antipsychotic dosage is reduced or stopped, i.e. tolerance is not the norm, although it may occur in a minority of patients. Tolerance to antipsychotic-induced hyperprolactinaemia may be more frequent in children and adolescents. This is suggested by a report of trial data from children and adolescents, aged 5–15 years, with behavioural problems who were treated with risperidone. Serum prolactin levels rose and peaked during the first 1–2 months of treatment, but then declined steadily to values within or close to the normal range by 3–5 months.
Hyperprolactinaemia can cause a wide range of symptoms (table II) or be asymptomatic. Individuals vary greatly as to the serum prolactin level at which symptoms appear. Some symptoms result from a direct effect of prolactin on target tissues and others from prolactin disrupting the functioning of the hypothalamic-pituitary axis and leading to reduced levels of gonadal hormones, i.e. estrogen in women and testosterone in men. Symptoms often appear within a few weeks of commencing an antipsychotic or increasing its dosage, but occasionally they appear after long-term treatment with a stable dose of antipsychotic. When oral antipsychotics are discontinued, prolactin levels usually return to the normal range within a few days, although the process can take up to 3 weeks. When conventional antipsychotic depots are stopped serum prolactin levels can take up to 6 months to normalise. There is increasing evidence that chronic hyperprolactinaemia can cause initially ‘silent’ health problems. The evidence is strongest for bone loss, with a recent study reporting an association between prolactin-raising antipsychotics and hip fracture. Prolactin-raising antipsychotics may be associated with a small increase in the risk of female breast cancer.
Health professionals often fail to detect symptoms of antipsychotic-induced hyperprolactinaemia. This may reflect unfamiliarity with the symptoms, reluctance on behalf of patients and clinicians to discuss endocrine and sexual adverse effects, and the fact that these symptoms are not visibly stigmatising in the way that, for example, drug-induced weight gain and parkinsonism are. When patients are prescribed prolactin-raising antipsychotics clinicians should routinely inquire about symptoms that may indicate hyperprolactinaemia.
4. Sexual Dysfunction
Sexual dysfunction is common in those treated with antipsychotics. In a questionnaire survey, 43% of people treated with antipsychotic drugs reported sexual dysfunction. Subjectively, sexual dysfunction is rated as one of the most distressing of the antipsychotic adverse effects. Despite this it is often unrecognised by health professionals and has been called ‘the unspoken adverse effect of antipsychotics’. It can be specified according to the phase of the sexual response cycle that is affected, i.e. desire (libido), arousal, orgasm. More than one phase may be affected in an individual.
It is unclear if rates of sexual dysfunction differ between conventional and atypical antipsychotics or between individual atypical drugs. This reflects the lack of comparative studies, conflicting data and methodological issues that hinder interpretation of data. For example, in one study the prevalence of sexual dysfunction associated with quetiapine or risperidone was 11.7% based on spontaneous reports but increased to 32% when assessed using a semi-structured interview. Several scales exist to assess sexual functioning and this hampers comparisons between studies. Studies also need to account for individual variation in what constitutes normal sexual functioning and confounding factors that can affect sexual functioning. These include concurrent physical illness, co-prescribed medication, alcohol use, relationship difficulties and the symptoms of the underlying psychiatric illness.
Conventional and atypical antipsychotics can cause impairment of arousal and orgasm in both sexes.[85,86] In men, conventional antipsychotics can cause failure of ejaculation, reduced ejaculatory volume, ‘retrograde’ ejaculation, painful orgasm and, rarely, spontaneous ejaculation without sexual stimulation, as well as priapism. Case reports have noted priapism after patients were treated with clozapine,[91,92] risperidone and olanzapine,[94,95] and after an overdose of quetiapine.
Antipsychotics can cause sexual dysfunction through multiple mechanisms, including sedation, hyperprolactinaemia (which can cause sexual dysfunction directly and indirectly by causing secondary hypogonadism) and antagonism of α-adrenergic, dopaminergic, histaminic and muscarinic receptors.
Differential effects on serum prolactin levels and serotonin transmission provide theoretical reasons as to why atypical antipsychotics may cause less sexual dysfunction than conventional agents and why differences may exist between the atypical agents. Increased serotonergic neurotransmission, via 5-HT2A and 5-HT2C receptors, inhibits male and female sexual behaviour. This explains the sexual problems, predominantly delayed ejaculation in men and anorgasmia in women, seen with SSRIs. Conversely, in animal models, compounds that block 5-HT2 receptors (e.g. cyproheptadine) facilitate sexual behaviour. All atypical antipsychotics, with the exception of amisulpride, are 5-HT2 receptor antagonists, and, in theory, this may improve sexual functioning or at least mitigate the potential of the drug to cause sexual dysfunction.
Whether these mechanisms translate into clinical differences is unclear, although evidence suggests that prolactin is relevant. Ghadirian et al. found that elevated serum prolactin levels correlated with increasing sexual dysfunction in men, but not women, treated with conventional antipsychotics. Smith et al. found the opposite, with high prolactin levels being associated with sexual dysfunction, specifically problems of arousal, in women but not men treated with conventional antipsychotics. The multifactorial aetiology of sexual dysfunction may explain these contradictory results. At an individual level at least, hyperprolactinaemia can be relevant. Case reports[77,98] and one small study describe an improvement in sexual dysfunction, associated with antipsychotic-induced hyperprolactinaemia, following a switch to a prolactin-sparing antipsychotic and normalisation of serum prolactin levels.
Several studies have compared clozapine to other antipsychotics in terms of effect on sexual functioning. One study, comparing clozapine and haloperidol in men and women, found no difference in sexual functioning. A more recent study in men found that clozapine was associated with significantly better sexual functioning than conventional antipsychotics in terms of several domains, including frequency of orgasm and sexual satisfaction. A third study, comparing risperidone, clozapine and conventional antipsychotics in men, showed a decrease in sexual functioning in all three groups; however, the clozapine group showed significantly less reduction in sexual interest than the other two groups, and significantly less impairment of erections and orgasm than the risperidone group.
In a recent retrospective cross-sectional study, sexual dysfunction was less common with quetiapine (18.2%; mean dosage 360.5 mg/day) than with haloperidol (38.1%; 10.6 mg/day), olanzapine (35.3%; 13.5 mg/day) or risperidone (43.2%; 5.3 mg/day). In this study sexual dysfunction appeared to be dose related with haloperidol, risperidone and olanzapine. In a review of studies, erectile failure was significantly less common with amisulpride than risperidone (1% vs 5%). A Cochrane review concluded that abnormal ejaculation was less common with olanzapine than risperidone, but that other parameters of sexual functioning did not differ between the two drugs.
In summary, there is insufficient evidence to categorically state whether the atypical agents differ in their effect on sexual function. Existing studies tentatively suggest that sexual dysfunction may be more common with risperidone than with olanzapine, quetiapine or clozapine. This may reflect differential effects on serum prolactin. Irrespective of this explanation, in individual patients, antipsychotic-induced hyperprolactinaemia can cause sexual dysfunction, reversal of which can improve sexual functioning.[77,98,99] For a review of the differential diagnosis and management of sexual dysfunction caused by psychiatric drugs readers are referred to the review by Baldwin et al.
5. Weight Gain and Metabolic Abnormalities
5.1 Weight Gain
Surveys of patients with schizophrenia, and their relatives, report weight gain to be a common and troubling adverse effect.[105,106] Weight gain is stigmatising, reduces self-esteem and is associated with several diseases, including hypertension, type II diabetes mellitus, coronary heart disease, respiratory problems and some cancers.
Allison et al. conducted a meta-analysis to determine average weight change after 10 weeks on standard doses of antipsychotics. There were marked differences between individual conventional and atypical drugs. Among the conventional drugs, weight change ranged from a reduction of 0.39kg with molindone to a gain of 3.19kg with thioridazine. Among the atypical drugs, the largest weight increase was 4.45kg with clozapine and the smallest increase was 0.04kg with ziprasidone. There were insufficient data to evaluate quetiapine, amisulpride was excluded from the analysis, and aripiprazole did not feature as it was unlicensed at the time. In comparative trials, clozapine, olanzapine, risperidone, quetiapine, sertindole and zotepine have all been associated with more weight gain than haloperidol. Nevertheless, the difference in mean weight gain seen with specific atypical and conventional antipsychotics means that a simple comparison of the weight gain potential of ‘atypicals’ and ‘conventionals’ is fairly meaningless.
Narrative reviews[110,111] that have considered long- and short-term data have come to similar conclusions as the Allison et al. meta-analysis regarding the rank order of the atypical drugs. For example, Haddad concluded that mean weight gain was greatest with clozapine and olanzapine, and least with ziprasidone and aripiprazole. However, it is important to emphasise that weight change with all antipsychotics shows a marked individual variation that is obscured by considering mean weight change alone, i.e. whatever the antipsychotic, treatment will be associated with some patients losing weight, some maintaining their weight and some gaining weight. For these reasons it is helpful to consider categorical data on weight gain, e.g. the percentage of patients with an increase of >7% of bodyweight or the percentage of patients who gain 0–5kg, 5–10kg, etc.
Figure 1 shows the percentage of patients gaining >7% of baseline bodyweight in short-term placebo-controlled trials (3–8 weeks) in patients with schizophrenia (data from US labels). The rates for each drug are not directly comparable owing to differences between the trials in duration (3–8 weeks) and patient characteristics, but rather it is the difference between each drug and the corresponding placebo group that matters. Two conclusions are apparent. First, each drug in figure 1 is associated with more than twice as many patients gaining >7% of baseline weight than placebo, i.e. all the drugs can cause significant weight gain in some patients. Second, the likelihood of weight gain varies markedly, being highest with olanzapine and lowest with aripiprazole and ziprasidone.
In phase I of the CATIE study significantly more patients treated with olanzapine gained >7% baseline weight than with other drugs (30% vs 7–16%), and the discontinuation rate due to weight gain and metabolic effects was significantly higher with olanzapine than with other drugs (9% vs 1–4%).
Predictors of weight gain include a good clinical response to treatment, low body mass index and, in the case of olanzapine, rapid weight gain and appetite increase during the first 6 weeks of treatment. The association between weight gain and antipsychotic response is influenced by methodological issues, and the magnitude of the effect remains debated. Antipsychotic dose, at least within the licensed therapeutic dosage range, does not have a significant effect on weight gain.
Clozapine is the best studied atypical antipsychotic in terms of long-term weight gain. Two studies reported steady weight gain throughout the first year of clozapine treatment, with 58% and 70% gaining >10% of bodyweight after 1 year. A 5-year prospective study reported a steady increase in bodyweight over the first 4 years of treatment, with weight remaining steady in the fifth year. Preliminary data from the naturalistic SOHO (Schizophrenia Outpatients Health Outcomes) study showed that mean weight with olanzapine, risperidone and all other drug cohorts, increased continuously over the first 3 years of treatment. In contrast, some data suggest that weight gain with olanzapine and risperidone may reach a plateau early on in treatment. A retrospective analysis of patients treated with olanzapine for a median duration of 2.54 years showed that mean weight gain stabilised at 39 weeks, with a last observation carried forward mean weight change of 6.26kg. Weight gain with risperidone has been reported to stabilise after as little as 10 weeks. Data on long-term weight gain with quetiapine are conflicting; some data suggest that weight gain continues over the first year of treatment, with a mean weight gain of 5.6kg, although other data suggest minimal long-term weight gain. Despite these contradictory results it is apparent that, in comparison to other atypical agents, clozapine and olanzapine are particularly likely to be associated with an early and rapid increase in weight.
In summary, mean weight gain is highest with clozapine and olanzapine, and approximately zero with aripiprazole and ziprasidone. However, all atypical drugs can cause significant (>7%) weight gain in some patients, although the relative risk varies markedly. Weight gain, unlike most antipsychotic adverse effects, does not appear to be influenced by changing drug dose within the therapeutic range. It can be influenced, at least to some extent, by lifestyle changes in which patients can actively participate.[119,120]
In the mid-1950s, hyperglycaemia and diabetes mellitus were reported as adverse effects of chlorpromazine.[121,122] Case reports and ADR reports of the ‘challenge-rechallenge’ type provide strong anecdotal evidence that atypical antipsychotics can impair glycaemic control. In these reports glycaemic abnormalities tend to appear soon after an atypical antipsychotic is started, resolve when it is stopped and re-occur when the atypical is re-introduced. Such cases have been reported for clozapine, olanzapine, risperidone and quetiapine. In summary, there is no doubt that atypical antipsychotics can (i) worsen glycaemic control in patients with pre-existing diabetes mellitus; and (ii) cause a range of glycaemic abnormalities, ranging from hyperglycaemia to frank type II diabetes mellitus, in patients without previously diagnosed diabetes mellitus. What is less clear is how often these adverse effects occur and to what extent the incidence differs between antipsychotics.
This is a controversial area with different opinions being expressed, even among expert consensus statements. A US consensus statement concluded that atypical antipsychotics differed in their propensity to cause hyperglycaemia, with clozapine and olanzapine showing the greatest effects, risperidone and quetiapine having intermediate effects, and aripiprazole and ziprasidone having lower effects. In contrast, European and Australian consensus statements did not recognise any difference among atypical agents in the risk of hyperglycaemia. Since these consensus statements were published, the CATIE results have become available and have helped to clarify the situation. The CATIE study demonstrated, in a double-blind design, that the atypical antipsychotics differ in their risk of causing hyperglycaemia.
Weight gain is a major risk factor for diabetes mellitus. Consequently, in the long-term one would expect antipsychotics that are associated with greater weight gain to be associated with higher rates of diabetes mellitus, assuming that all other risk factors were equivalent. Antipsychotics can also exert effects on glucose tolerance independent of weight gain. When modified oral glucose tolerance tests were performed in schizophrenic patients matched for adiposity and age, clozapine and olanzapine, but not risperidone, were associated with significant plasma glucose elevations compared with patients receiving typical antipsychotics. In a small prospective study, 11 (55%) of 20 patients developed abnormal glucose control within 4 months of starting clozapine. The impairment of glucose control was independent of changes in body mass index and insulin resistance.
Retrospective studies, including pharmacoepidemiological database studies, cannot account for some potential risk factors for diabetes mellitus and this limits their ability to identify the extent of any drug effects. Diabetic risk factors include age, race, weight, diet, smoking and exercise levels. Two studies[133,134] have reported insulin resistance in antipsychotic-naive patients, leading to the suggestion that schizophrenia may be an independent risk factor for hyperglycaemia, although other studies[135,136] have failed to replicate this. A high proportion of cases of diabetes mellitus within psychiatric populations are undiagnosed[137,138] and this leads to another weakness of pharmacoepidemiological studies, namely that they assume that screening for diabetes mellitus, and its diagnosis, are uniform across medication groups. Given these methodological issues, it is not surprising that pharmacoepidemiological studies have produced contradictory results regarding the relative risk of diabetes mellitus within the atypical group. Nevertheless, most pharmacoepidemiological studies report a higher incident risk of diabetes mellitus in patients treated with atypical compared with conventional antipsychotics. Whether this reflects a pharmacological difference in risk, a screening bias or some other effects is unclear.
One of the methodologically stronger pharmacoepidemiological studies is that conducted by Lambert et al., which compared the risk of new-onset type 2 diabetes mellitus following the initiation of atypical antipsychotics versus haloperidol. After adjusting for confounders, diabetes risk was increased equally with new use of olanzapine, risperidone or quetiapine (hazard ratio varied from 1.60 to 1.67). The investigators concluded that if the observed associations were causal, then approximately one-third of new cases of diabetes mellitus in patients taking olanzapine, risperidone and quetiapine may be attributed to these drugs.
To date, relatively few published trials have assessed glycaemic control in patients randomised to an atypical antipsychotic versus another antipsychotic.[15,140,141] Lindenmayer et al. described a cohort of patients with treatment-resistant schizophrenia randomised to clozapine, haloperidol, olanzapine or risperidone for 14 weeks. In the latter part of the study, above maximum licensed dosages of antipsychotic medication were employed. Diabetes mellitus was defined as a fasting blood glucose level of >7.0 mmol/L. In all three atypical cohorts there was a high incidence of ‘diabetes’ (ranging from 14% to 21%). The rates did not differ between the three atypical groups, although these were all higher than the 4% incidence seen in the haloperidol cohort. The high rates of ‘diabetes’ in the atypical cohorts may reflect the fact that patients had treatment-resistant schizophrenia and were treated with high dosages of medication.
Lieberman et al. reported a 12-month RCT in which 160 patients with first-episode schizophrenia in China were randomised to clozapine or chlorpromazine. Mean fasting blood glucose levels did not differ between the two groups at baseline or 12 months, the 12-month mean glucose levels were within normal limits, and there were no incident cases of diabetes mellitus in either group. However, this is not an ideal study on which to draw conclusions about the hyperglycaemic risk of antipsychotics as the subjects were young, would have a large β-cell insulin reserve and were followed for only 12 months. Furthermore, diabetes mellitus has a low prevalence in China, which may reflect genetic factors and a relative absence of lifestyle risk factors. Consequently, the propensity of antipsychotics to cause diabetes mellitus may be less likely to manifest in this population.
In the CATIE study, olanzapine was associated with a significantly greater increase in glycosylated haemoglobin than the other drugs, but there was no significant difference between treatment groups in the degree of change in mean blood glucose levels or in the number of patients who required the addition of an oral glucose-lowering agent. Interpretation is confounded by the inclusion of fasting and nonfasting blood samples and the high proportion of patients in each drug group with pre-existing diabetes mellitus at baseline (9–11%).
In summary, both conventional and atypical antipsychotics can worsen glycaemic control. An overview of the existing data indicates that clozapine and olanzapine have a higher risk of causing hyperglycaemia than the other atypical antipsychotics and, where comparator studies exist, conventional antipsychotics, including haloperidol and perphenazine. Further research is required to investigate this area.
Irrespective of antipsychotic medication, patients with schizophrenia are a high-risk group for diabetes mellitus by virtue of their lifestyle in terms of diet, weight and exercise. Various consensus statements and expert reviews recommend that blood glucose levels are monitored during antipsychotic treatment.[128, 129, 130,142] Measuring blood glucose levels prior to commencing antipsychotic treatment, after 4 months of treatment, and annually thereafter has been suggested as a suitable level of monitoring. Clinicians need to be alert to symptoms that may indicate the appearance of diabetes mellitus (e.g. lethargy, polyuria, polydipsia, weight loss, recurrent infections) and to the possibility that the commencement or switching of an antipsychotic may alter glycaemic control in those patients with pre-existing diabetes mellitus.
Serum lipid levels above the upper limit of normal are common in the general population. In a cross-sectional survey of the US population conducted between 1999 and 2000, over half of the Caucasians surveyed, aged between 29 and 74 years, had raised serum levels of low-density lipoprotein-cholesterol (LDL-C) and total cholesterol. The first suggestion that antipsychotics may elevate serum lipid levels came in 1970, but clarifying their effect has proved difficult because of a lack of long-term prospective studies with systematic assessment of fasting lipid levels. Much current data comes from cross-sectional or retrospective analyses. Other weaknesses include inadequate reporting of lipid fractions, missing samples (raising the question of whether or not the available data are representative) and the fact that among the atypical antipsychotics there are little lipid data available for amisulpride or zotepine.
In the CATIE study, olanzapine was associated with significantly greater increases in serum levels of cholesterol and triglycerides than the other drugs, even after adjustment for treatment duration. Ziprasidone and risperidone were the only drugs in the CATIE study to be associated with reductions in mean serum levels of cholesterol and triglycerides. In a recent literature review, Melkersson and Dahl concluded that the relative risk for hyperlipidaemia was highest for clozapine and olanzapine, moderate for quetiapine, and low for risperidone and ziprasidone. Aripiprazole also has a low risk of causing dyslipidaemia. In a 26-week RCT that compared olanzapine and aripiprazole, changes in fasting plasma levels of total cholesterol, high-density lipoprotein-cholesterol (HDL-C) and triglycerides were significantly different in the two treatment groups, with a worsening of the lipid profile seen in the olanzapine-treated group. At endpoint there was a mean weight loss of 1.37kg with aripiprazole compared with a mean increase of 4.23kg with olanzapine, among patients who remained on medication (p < 0.001). A retrospective chart review of patients who switched to aripiprazole from other atypical drugs showed a decrease in weight, serum levels of total cholesterol and LDL-C.
It seems that the relative risk of dyslipidaemia among the atypical drugs mirrors their potential to cause weight gain (see section 5.1). Antipsychotic-induced elevation of triglycerides may partly reflect weight gain, rather than being a totally independent effect. It is unclear if serum triglyceride levels normalise once weight stabilises.
Data are limited regarding the effect of conventional antipsychotics on lipid levels, but high-potency drugs (e.g. haloperidol) appear to carry a lower risk of hyperlipidaemia than low-potency drugs (e.g. chlorpromazine and thioridazine).
A survey examining the knowledge of psychiatrists and their practice patterns found that only 22% recognised dyslipidaemia as a possible complication of atypical antipsychotic therapy. Various guidelines and consensus statements recommend that serum lipid levels be monitored during antipsychotic treatment.[127,149,151]
6. Cardiovascular Adverse Effects
6.1 Postural Hypotension
The propensity of antipsychotics to cause postural or orthostatic hypotension is correlated with the degree of antagonism at α1-adrenoceptors. Orthostatic hypotension is more common in older patients, those with cardiovascular disease and those receiving antihypertensive drugs. It is often manageable with careful dose adjustment and patients frequently become partially or fully tolerant to this adverse effect.
Clozapine is the atypical drug most often associated with postural hypotension. Risperidone and quetiapine can cause postural hypotension in some patients, particularly with rapid titration.
6.2 QTc Prolongation
A possible association between antipsychotics and sudden unexpected death was raised by Reinert and Hermann in 1960. Soon after repolarisation, abnormalities were noted in a substantial proportion of phenothiazine-treated patients, suggesting an underlying cardiac mechanism. There is now a general consensus that at least some antipsychotic drugs increase the risk of arrhythmias and sudden unexpected death, and that prolongation of the QTc interval is a marker of such risk. Controversy remains regarding the degree of risk and how one should balance this against the benefits of antipsychotic treatment.
These issues and uncertainties are not unique to antipsychotics, but apply to a wide range of drugs that prolong the QTc interval, including certain antibacterials, antihistamines and cardiac drugs. All these drugs increase the QTc interval by blocking the delayed rectifier potassium channel (IKR) in the cell membrane of cardiac myocytes, thereby slowing the outward movement of intracellular potassium and prolonging cardiac repolarisation. This increases the risk of ‘torsade de pointes’, an arrhythmia that can be self-limiting, cause palpitations, syncope and dizziness, or progress to ventricular fibrillation and sudden death. The QTc interval is a measure of ventricular repolarisation and acts as a marker of arrhythmic risk. There is a consensus that a QTc interval of >500ms, or an absolute increase of 60ms compared with drug-free baseline, puts a patient at significant risk of torsade de pointes, ventricular fibrillation and sudden death.
QTc prolongation with antipsychotics, as well as other drugs, is dose related. Antipsychotics associated with a greater risk of QTc prolongation include pimozide and thioridazine among the conventional antipsychotics, and ziprasidone and sertindole among the atypical drugs.
Sertindole was licensed in the UK in 1996 for the treatment of schizophrenia. It was known to cause QTc prolongation, and warnings related to this were included in the prescribing information. It was voluntarily withdrawn by the manufacturer in 1998 following ADR reports of sudden cardiac death and serious but nonfatal cardiac arrhythmia. It is currently only available direct from the manufacturer for patients in specific clinical studies. A Cochrane review identified two trials that compared sertindole to haloperidol. In both trials only sertindole was associated with a QTc of ≥500ms. In a small prospective study of patients prescribed amisulpride, olanzapine, sertindole and clozapine only those prescribed sertindole (12 mg/day) showed a significant increase in the QTc interval.
Pfizer study 054 compared the cardiac effect of ziprasidone, thioridazine, quetiapine, risperidone and olanzapine, and measured the QTc interval at the time of peak plasma concentration. The greatest effect on the QTc interval (Bazett correction), compared with drug-free baseline, was seen with thioridazine (150mg twice daily; QTc increased by 35.6ms) and ziprasidone (80mg twice daily; QTc increased by 20.3ms), which were the only agents to be associated with a QTc increase of ≥75ms (figure 2). The US label for ziprasidone contains a warning of the potential for QTc prolongation and sudden death.
Where QTc-prolonging drugs have been linked to sudden death they have often acted in conjunction with other risk factors for arrhythmias (table III). Of particular importance is the co-prescription of another QTc-prolonging drug or a drug that inhibits the metabolism of the QTc-prolonging drug, leading to higher plasma concentrations.
In summary, QTc prolongation can occur with all atypical antipsychotics. With the exception of ziprasidone and sertindole, the effect is usually minor if the atypical drug is prescribed within the licensed dosage range and the patient does not have other risk factors for arrhythmias. If either of these safeguards is absent then extra vigilance and monitoring is advisable, for example ECG monitoring, avoiding co-prescription of medications that could increase arrhythmic risk and checking plasma electrolyte levels. A detailed review of these issues is provided by Haddad and Anderson.
6.3 Other Cardiac Effects
There have been rare reports of myocarditis, cardiomyopathy, pericarditis and pericardial effusions in association with clozapine treatment. Estimates of the incidence of myocarditis vary between 1 in 500 and 1 in 10 000. Postmarketing experience suggests that the risk of myocarditis is highest in the first 2 months of treatment, whereas cardiomyopathy generally occurs later. Some cases of myocarditis and cardiomyopathy have been fatal. Because of these risks, physical examination and a careful history, inquiring about cardiac symptoms and disease, is recommended before starting clozapine therapy. Clozapine should be discontinued if a patient presents with symptoms that suggest myocarditis or cardiomyopathy, i.e. new-onset fatigue, dyspnoea, fever, chest pain, palpitations, ST-segment abnormalities or arrhythmias. An urgent diagnostic evaluation should be obtained from a cardiologist. Patients with clozapine-induced myocarditis or cardiomyopathy should not be re-exposed to clozapine.
Six cases of pulmonary embolism and six of venous thrombosis were reported in Sweden between 1989 and 2000 in patients prescribed clozapine; however, a conclusive causal relation could not be proven.
Sinus tachycardia can occur with all atypical drugs. It is most common with clozapine, where it has been reported to occur in about one-quarter of users, particularly during dose titration in early treatment. Tachycardia with clozapine is usually benign and can be reduced with slower titration and/or β-adrenoceptor antagonists treatment if this symptom persists; however, tachycardia can also be a key symptom of myocardial disease. Consequently, clozapine-treated patients who have persistent tachycardia at rest, particularly in the first 2 months of treatment, should be closely monitored for signs and symptoms of myocarditis/cardiomyopathy.
Other ECG findings reported with various atypical agents include palpitations, premature atrial and ventricular beats, ST segment depression and atrioventricular block. Hypertension can occur with antipsychotics, particularly clozapine, but is far less common than hypotension (see section 6.1).
7. Antimuscarinic Adverse Effects
Antimuscarinic adverse effects (also termed anticholinergic adverse effects) include dry mouth, blurred vision, urinary retention, constipation and, in severe cases, cognitive impairment and delirium. Dry mouth may contribute to dental decay, blurred vision can cause falls, and constipation can lead to gastrointestinal obstruction.
Many conventional antipsychotic drugs, particularly thioridazine, can cause antimuscarinic adverse effects. Among the atypical antipsychotics such effects are most pronounced with clozapine. This reflects the relatively high affinity of clozapine for muscarinic receptors, although in radioligand binding studies this is considerably lower than that of classical muscarinic antagonists. In a model that assessed the anticholinergic activity of atypical antipsychotics at therapeutic doses, clozapine, olanzapine and quetiapine all showed dose-dependent increases in anticholinergic activity. In contrast, aripriprazole, risperidone and ziprasidone did not demonstrate anticholinergic activity at any of the concentrations studied.
In a comparative study, in which concomitant anticholinergic drugs were excluded, anticholinergic symptoms, with the exception of dry mouth (i.e. constipation, urinary disturbances and tachycardia/palpitations), were more marked in clozapine-treated than in olanzapine-treated patients. Neither group showed any global cognitive deficits. Consistent with the symptomatic differences, clozapine-treated patients had serum anticholinergic levels (assessed using a radio-receptor binding assay) approximately 5-fold higher than olanzapine-treated patients.
Paradoxically, clozapine causes hypersalivation, not dry mouth. The aetiology of hypersalivation with clozapine is unclear. Clozapine may increase salivation through blockade of α2-adrenoceptors and/or through activation of muscarinic M4 receptors in the salivary glands. An alternative hypothesis is that saliva flow is not increased and that clozapine interferes with swallowing, causing pooling of saliva in the mouth.
Constipation with clozapine is often persistent and can lead to obstruction, which may occasionally be fatal. Gastro-oesophageal reflux is common with clozapine and may reflect the antimuscarinic effects of the drug, leading to relaxation of the cardiac sphincter, delayed gastric motility and delayed transit time.[170,171] Due its anticholinergic properties, caution and careful supervision are required if clozapine is prescribed to patients with prostatic enlargement or narrow-angle glaucoma.
8. Miscellaneous Adverse Effects
8.1 Blood Dyscrasias
The most common dyscrasia seen with the atypical antipsychotics is neutropenia with clozapine. Neutropenia is defined as a neutrophil count <1500 cells/mm3. When the neutrophil count falls <1000 cells/mm3 there is a significant increase in the risk of infection. Agranulocytosis is a severe state of neutropenia when the absolute neutrophil count falls <500 cells/mm3. It can be asymptomatic or manifest with a range of symptoms, including fever, headache, sore throat, stomatitis, diarrhoea, myalgia, arthralgia and urinary frequency.
Clozapine is associated with agranulocytosis in about 0.8% of patients. More than 84% of such cases occur within 3 months of starting treatment and over 90% within the first 6 months of treatment. Agranulocytosis is more than twice as frequent in Asians as in Caucasians. The risk increases with age[11,172] and is higher among women. Differences in the risk factors for clozapine-induced agranulocytosis and neutropenia imply that each disorder may have a distinct mechanism. Regular haematological monitoring allows clozapine to be stopped if parameters drop below one of two set thresholds, i.e. a white cell count (WCC) <3000 cells/mm3 or an absolute neutrophil count <1500 cells/mm3. With such safeguards, mortality is exceptionally low. With the other atypical drugs, dyscrasias are rare, and routine haematological monitoring is not required.
Benign reductions in neutrophil counts are more common in Afro-Caribbeans than in Caucasians, and will account for some cases of neutropenia seen during clozapine treatment. Distinguishing between benign ethnic neutropenia and clozapine-induced neutropenia is not possible with absolute certainty, although certain factors can help differentiation. A relatively high pretreatment WCC and a precipitous fall in the WCC (over 1–2 weeks or less) soon after starting clozapine would all suggest clozapine-induced neutropenia. Reports have described using lithium to elevate the WCC in those with benign ethnic neutropenia prior to starting clozapine and neutropenia that has developed during clozapine treatment.[175,176] Great care is required if this strategy is used. We do not recommend that lithium be used to treat neutropenia that is believed to be due to clozapine. Lithium does not protect against agranulocytosis and it must be remembered that clozapine-induced neutropenia can occur in those with benign ethnic neutropenia.
8.2 Discontinuation Symptoms
Antipsychotic discontinuation (withdrawal) symptoms are much neglected. The first report was in 1960 when a group of patients, without psychosis, were noted to develop restlessness, insomnia, nausea and vomiting after stopping chlorpromazine, with the symptoms resolving on restarting chlorpromazine. Since then motor and nonmotor discontinuation symptoms have been described with several antipsychotics. Clozapine is the only atypical agent with well documented discontinuation symptoms. However, since many drugs acting on the CNS cause discontinuation symptoms,[24,179] it would be surprising if other atypical antipsychotics did not cause such effects, at least in some patients.
Clozapine discontinuation symptoms usually commence within 2–3 days after stopping the drug, and can include anxiety, insomnia, motor restlessness, confusion, altered consciousness, nausea and diaphoresis.[180, 181, 182] Cases of withdrawal dystonias, dyskinesias and tics have been described.[183,184] Some patients develop a rapid psychosis shortly after clozapine withdrawal.[181,182] Whether this represents re-emergence of the underlying psychosis or a discontinuation effect is unclear.
8.3 Neuroleptic Malignant Syndrome
The neuroleptic malignant syndrome (NMS) is an uncommon but potentially life-threatening condition characterised by muscle rigidity, autonomic instability, confusion and fever. Various sets of diagnostic criteria for NMS have been proposed, including those outlined in DSM-IV, i.e. severe muscle rigidity and elevated temperature associated with the use of an antipsychotic drug; and two or more of the following features: diaphoresis, dysphagia, tremor, incontinence, altered level of consciousness, mutism, tacycardia, elevated or labile blood pressure, leukocytosis and laboratory evidence of muscle injury (e.g. elevated creatine phosphokinase).
NMS has been attributed to blockade of D2 receptors in the hypothalamus and striatum, but the pathophysiology remains uncertain. Most cases involve haloperidol but other antipsychotic drugs[186,187] can cause the syndrome, as can various non-antipsychotic drugs. A review identified 68 published cases of NMS (21 females and 47 males) associated with atypical antipsychotic drugs (clozapine = 21, risperidone = 23, olanzapine = 19; quetiapine = 5). NMS has also been reported with amisulpride and aripiprazole. The diagnosis of NMS in patients receiving clozapine is complicated by the benign fever and tachycardia commonly encountered during the early stage of treatment with clozapine.
Estimates of the incidence of NMS vary widely, partly reflecting the diagnostic criteria adopted, but overall the incidence appears to have decreased. This may reflect decreased use of antipsychotic polypharmacy, decreased use of high-dose high-potency agents, and the increasing use of atypical antipsychotics.
Sedation with antipsychotics tends to be dose related and patients often become tolerant to the effect. With the conventional antipsychotics, sedation is more marked with chlorpromazine than haloperidol. Among the atypical agents it is most pronounced with clozapine, where it can be dose limiting and a cause of poor compliance.[193,194] Clinical experience suggests that, clozapine apart, olanzapine and quetiapine are the most sedating of the atypical drugs and aripiprazole the least sedating.
All antipsychotics can lower the seizure threshold and should therefore be used with caution in patients with a history of seizures and those with organic brain disorders, such as Alzheimer’s disease, that potentially lower the seizure threshold. As a general rule, the more sedating the antipsychotic the more likely it is to cause seizures. Chlorpromazine appears to be the most epileptogenic of the conventional antipsychotics. The atypical antipsychotic most commonly linked with seizures is clozapine. The association is dose related; for dosages <300 mg/day the seizure rate is about 1%, while for dosages >600 mg/day the rate is 4.4%. The onset of seizures with clozapine may be preceded by severe myoclonus. Of the remaining atypical drugs, seizures appear most common with zotepine. With other atypical drugs seizures are uncommon. Depot antipsychotics should be avoided in patients with epilepsy, not because they are of high epileptogenic risk, but because if seizures occur then the antipsychotic cannot be quickly withdrawn.
8.6 Cerebrovascular Events
Antipsychotics are frequently used to treat behavioural and psychiatric symptoms of dementia, although none are licensed for this indication. In 2004 the Committee on Safety of Medicines (CSM) in the UK issued recommendations regarding the use of olanzapine and risperidone in patients with dementia. This followed a review of two meta-analyses showing that elderly patients with dementia treated with these drugs had an approximately 3-fold increased risk of cerebrovascular events (including strokes and transient ischaemic attacks) compared with patients who received placebo. The CMS recommendations included the following:
Olanzapine and risperidone should not be used in the treatment of behavioural symptoms of dementia as the risk of stroke outweighed the benefits of medication in this situation.
The use of risperidone to treat acute psychotic conditions in elderly patients with dementia should be limited to the short-term and should be under specialist advice (olanzapine is not licensed for management of acute psychosis).
Prescribers should carefully consider the risk of cerebrovascular events before treating any patient with a previous history of stroke or transient ischaemic attack. Consideration should be given to other risk factors for cerebrovascular disease.
The CSM concluded that there was insufficient information to include other antipsychotics in these recommendations, but that an increased risk of stroke could not be excluded with other antipsychotics. Since that time several other studies have investigated this possibility.
A meta-analysis of placebo-controlled trials of atypical antipsychotic drugs (aripiprazole, olanzapine, quetiapine, risperidone) in patients with dementia showed a small but significant increase in mortality in those randomised to drugs (3.5%) versus placebo (2.3%), and also a significantly increased risk for cerebrovascular events. There was no evidence of a differential mortality risk for individual drugs. A recent large RCT found that the adverse effects offset advantages in the efficacy of atypical antipsychotic drugs for the treatment of psychosis, aggression or agitation in patients with Alzheimer’s disease.
Three cohort studies have compared the risk of death or stoke in elderly patients (aged >65 years) treated with conventional versus atypical antipsychotics. Two studies found no difference in the risk of stroke between those treated with atypical and conventional antipsychotics.[203,204] In the third study, conventional antipsychotics were associated with a significantly higher adjusted risk of death than atypical antipsychotics. A case-control study of residents of US nursing homes with a diagnosis of dementia assessed the risk of being hospitalised for a cerebrovascular event. After controlling for potential confounders, no increased risk of stoke was seen in those treated with either atypical or conventional antipsychotics.
In summary, in several meta-analyses, atypical antipsychotics are associated with a higher risk of cerebrovascular events and/or death than placebo in elderly patients with dementia.[199, 200, 201] At present there is no evidence of a differential risk between individual atypical antipsychotics regarding these adverse effects. In the US, the labels of all atypical drugs carry a warning that elderly patients with dementia-related psychosis treated with atypical antipsychotics are at an increased risk of death compared with those receiving placebo. The lack of placebo-controlled trials of conventional antipsychotics in elderly patients with dementia cannot be interpreted as indicating that these drugs are without such risks. Indeed, cohort studies suggest that the risk of stoke or death is equal, if not higher, with conventional antipsychotics than with atypical drugs. We recommend that clinicians consider these risks and the potential benefits of antipsychotics, on an individual patient basis when managing noncognitive symptoms of dementia. Where appropriate, nonpharmacological interventions should be tried first and atypical drugs reserved for the treatment of severe symptoms, particularly if associated with concerns about risk and patient distress. Data comparing atypical and conventional antipsychotics suggest that conventional drugs should not be automatically used to replace atypical agents in response to these safety concerns.
9. Conclusions and Clinical Implications
Adverse effects are important as they can impair quality of life, cause stigma, result in morbidity and mortality, and contribute to poor adherence with medication, which may lead to an increased relapse rate. Knowledge of the relative risk of adverse effects among the atypical antipsychotics facilitates informed prescribing decisions by the clinician and patient, and should minimise the burden of adverse effects. Our review indicates markedly different risks between the atypical drugs for a number of adverse effects. Table IV summarises these risks. It must be emphasised that these relative risks are approximate and are based on populations treated with each drug; individuals may differ markedly in their susceptibility to any adverse effect.
Four main conclusions can be drawn from this review:
A comprehensive assessment of tolerability requires a review of all the relevant data and not simply RCTs. Other data that should be considered includes observational studies and postmarketing surveillance. This is because each source of data has its strengths and weaknesses.
The variation in tolerability between the atypical antipsychotics means that it is misleading to regard them as a uniform drug ‘class’ and that the term ‘atypical’ has only limited usefulness. This view is also supported by differences in their mechanism of action and their efficacy.
Broad statements comparing the relative risk of specific adverse effects between ‘atypical’ and ‘conventional’ antipsychotics are largely meaningless; rather, comparisons should be made between specific atypical and specific conventional antipsychotics.
Adverse effects are usually dose dependent and patient characteristics, including age and gender, are important risk factors. The influence of these factors should be considered in clinical practice and in the interpretation of research data.
Selection of an antipsychotic should always be made on an individual patient basis. Where possible, patients should be involved in prescribing decisions and this should involve discussion about adverse effects. The degree to which this occurs will depend on the desire of the patient to be involved in treatment decisions, as well as their mental state. If a patient is too unwell to be involved in a detailed discussion at the time of prescribing, further discussion should take place once their mental state allows. A collaborative approach to prescribing between the patient and clinician can lead to greater patient satisfaction with treatment and higher rates of medication adherence. Tolerability is only one of several factors that the patient and clinician should consider when selecting an antipsychotic; others include efficacy, range of formulations (tablet, oral-dispersible tablet, liquid and long-acting injection) and previous response of the patient to medication.
As most adverse effects are dose related, treatment should use the minimal effective dose of medication, particularly in groups vulnerable to adverse effects, such as the elderly and children and adolescents. Unnecessary polypharmacy should be avoided. Regular inquiry as to possible adverse effects is an essential part of clinical follow-up; it is not sufficient to simply rely on patients volunteering information regarding adverse effects they may be experiencing. Plasma glucose and lipid levels should be monitored during treatment to detect hyperglycaemia and hyperlipidaemia.
The terms ‘atypical antipsychotic’ and ‘conventional antipsychotic’ are used throughout this review and are synonymous with the terms ‘second-generation’ and ‘first-generation’ antipsychotic, respectively.
The authors wish to thank Dr Bob Barber, Consultant Psychiatrist for the Elderly, Newcastle-upon-Tyne, England, for his helpful comments on the section regarding cerebrovascular events.
Peter Haddad has received fees for lecturing and/or consultancy from AstraZeneca, Bristol-Myers Squibb, Lilly, Janssen-Cilag and Novartis. Sonu Sharma has no conflicts of interest that are directly relevant to the content of this review. No sources of funding were used to assist in the preparation of this review.
- 7.Murray CJ, Lopez AD. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 projected to 2020. Cambridge (MA): Harvard School of Public Health, 1996Google Scholar
- 8.Haddad P, Dursun S, Deakin B, editors. Adverse syndromes and psychiatric drugs: a clinical guide. Oxford: Oxford University Press, 2004Google Scholar
- 14.British Medical Association and Royal Pharmaceutical Society of Great Britain. British national formulary. 53rd ed. London: BMJ Publishing Group Ltd and RPS Publishing, 2007 MarGoogle Scholar
- 20.Casey DE. Neuroleptic induced extrapyramidal syndromes and tardive dyskinesia. In: Hirsch S, editor. Schizophrenia. Oxford: Blackwell, 1995: 546–55Google Scholar
- 23.Prior TI, Baker GB. Anticholinergic toxicity. In: Haddad P, Dursun S, Deakin B, editors. Adverse syndromes and psychiatric drugs. Oxford: Oxford University Press, 2004: 293–7Google Scholar
- 24.Ashton H. Benzodiazepine dependence. In: Haddad P, Dursun S, Deakin B, editors. Adverse syndromes and psychiatric drugs. Oxford: Oxford University Press, 2004: 239–59Google Scholar
- 28.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 (CUt-LASS 1). Arch Gen Psychiatry 2006; 63(10): 1079–87PubMedCrossRefGoogle Scholar
- 30.National Institute for Clinical Excellence. Core interventions in the treatment and management of schizophrenia in primary and secondary care. Clinical guideline 1. London: National Institute for Clinical Excellence, 2002 DecGoogle Scholar
- 32.Marder SR, Miebach RC. Risperidone in the treatment of schizophrenia. Am J Psych 1994; 151: 825–35Google Scholar
- 35.Jayaram MB, Hosalli P, Stroup S. Risperidone versus olanzapine for schizophrenia. Cochrane Database Syst Rev 2006 Apr 19; (2): CD005237Google Scholar
- 36.Gerlach J. Improving outcome in schizophrenia: the potential importance of EPS and neuroleptic dysphoria. Ann Clin Psych 2002; 14(1): 47–57Google Scholar
- 41.Kane JM, Jeste DV, Barnes TRE, et al. Tardive dyskinesia a task force report of the American Psychiatric Association. Amlington (VA): American Psychiatric Association, 1992Google Scholar
- 43.Kane JM. Tardive dyskinesia in affective disorders. J Clin Psychiatry 1999; 60 Suppl. 5: 43–7, discussion 48-9Google Scholar
- 46.Beasley CM, Dellva MA, Tamura RN, et al. Randomised double blind comparison of the incidence of tardive dyskinea in patients in patients with schizophrenia during long term treatment with olanzapine or haloperidol. Br J Psych 1999; 175: 391–2Google Scholar
- 47.Csernansky JG, Mahmoud R, Brenner R. A comparison of risperidone and haloperidol for the prevention of relapse in patients with schizophrenia. Risperidone-USA-79 Study Group. N Engl J Med 2002; 346(1): 16–22Google Scholar
- 65.Tollefson GD, Birkett MA, Kiesler GM, et al. Double-blind comparison of olanzapine versus clozapine in schizophrenic patients clinically eligible for treatment with clozapine. Lilly Resistant Schizophrenia Study Group. Biol Psychiatry 2001; 49(1): 52–63Google Scholar
- 80.Hellewell JSE. Antipsychotic tolerability: the attitudes and perceptions of medical professionals, patients and caregivers towards the side effects of antipsychotic therapy [abstract]. Eur Neuropsychopharmacol 1998; 8 Suppl.: S248Google Scholar
- 83.Peuskens J, Sienaert P, De Heft M. Sexual dysfunction: the unspoken side effect of antipsychotics. Eur Psychiatry 1998; 13Suppl. 1: S23–30Google Scholar
- 88.Baldwin DS, Birtwistle J. Schizophrenia, antipsychotic drugs and sexual function. Prim Care Psychiatry 1997; 3: 115–23Google Scholar
- 89.Keitner GI, Selub S. Spontaneous ejaculations and neuroleptics. J Clin Psychopharmacol 1993; 3: 34–6Google Scholar
- 93.Ernes CE, Millson RC. Risperidone-induced priapism. Can J Psychiatry 1994; 39: 315–6Google Scholar
- 102.Bobes J, Garc A-Portilla MP, Rejas J, et al. Frequency of sexual dysfunction and other reproductive side-effects in patients with schizophrenia treated with risperidone, olanzapine, quetiapine, or haloperidol: the results of the EIRE study. J Sex Marital Ther 2003; 29(2): 125–47PubMedCrossRefGoogle Scholar
- 104.Baldwin DS, Mayers AG, Lambert A. Sexual dysfunction. In: Haddad P, Dursun S, Deakin B, editors. Adverse syndromes and psychiatric drugs: a clinical guide. Oxford: Oxford University Press, 2004: 125–44Google Scholar
- 111.Haddad PM. Weight change with atypical antipsychotics in the treatment of schizophrenia. J Psychopharmacol 2005; 19(6 Suppl.): 17–27Google Scholar
- 115.Novick D, Haro JM, Suarez D, et al. Tolerability of outpatient antipsychotic treatment: 36-month results from the European Schizophrenia Outpatients Health Outcomes (SOHO) study [poster]. CINP Congress; 2006 Jul 9–13; Chicago (IL)Google Scholar
- 122.Cooperberg AA, Eidlow S. Haemolytic anaemia, jaundice and diabetes mellitus following chlorpromazine therapy. J Can Med Assoc 1956; 75: 746–9Google Scholar
- 129.Expert group. ‘Schizophrenia and diabetes 2003’ expert consensus meeting, Dublin 3–4 October 2003: consensus summary. Br J Psychiatry Suppl 2004; 47: S112–4Google Scholar
- 134.Cohn TA, Wolever T, Bois D, et al. First epsiode and neuroleptic free patients with schziophrenia have reduced insulin sensitivity: a minimal model analysis [abstract]. Schizophren Bull 2005; 31: 197–8Google Scholar
- 146.McQuade RD, Stock E, Marcus R, et al. A comparison of weight change during treatment with olanzapine or aripiprazole: results from a randomized, double-blind study. J Clin Psychiatry 2004; 65 Suppl. 18: 47–56Google Scholar
- 151.Expert Consensus Meeting. Metabolic and lifestyle issues and severe mental illness: new connections to well-being? J Psychopharmacol 2005; 19(6 Suppl.): 118–22Google Scholar
- 152.Physicians’ desk reference. 55th ed. Montvale (NJ): Medical Economics Company Inc, 2001Google Scholar
- 156.Committee for Proprietary Medicinal Products (CPMP). Points to consider: the assessment of the potential for QT interval prolongation by non-cardiovascular medicinal (CPMP/986/96) [online]. Available from URL: http://www.coresearch.biz/regulations/cpmp.pdf [Accessed 2007 Aug 22]
- 157.Lewis R, Bagnall AM, Leitner M. Sertindole for schizophrenia. Cochrane Database Syst Rev 2005; (3): CD001715Google Scholar
- 159.US FDA, Psychopharmacological Drugs Advisory Committee. Briefing document for Zeldox capsules (ziprasidone hydrochloride), 19 July, 2000 [online]. Available from URL: http://www.fda.gov/ohrms/dockets/ac/OO/backgrd/3619b1a.pdf [Accessed 2007 Sep 25]
- 160.Committee on Safety of Medicines and the Medicines Control Agency. Clozapine and cardiac safety: updated advice for prescribers. Curr Probl Pharmacovigilance 2002; 28: 8Google Scholar
- 173.Clozaril® summary of product characteristics [online]. Available from URL: http://emc.medicines.org.Uk/emc/assets/c/html/displaydoc.asp?documentid=1277 [Accessed 2007 Sept 11]
- 181.Borison RL, Diamond BI, Sinha D, et al. Clozapine withdrawal rebound psychosis. Psychopharmacology Bull 1998; 24: 260–3Google Scholar
- 185.American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association, 1994Google Scholar
- 192.Keck PE, Pope HG. Declining frequency of neroleptic malignant syndrome. Am J Psych 1991; 148: 880–2Google Scholar
- 193.Kane JM. Factors which can make patients difficult to treat. Br J Psych Suppl 1996; 31: 10–4Google Scholar
- 194.Daniel DG, Goldberg TE, Weinberger DR, et al. Different side effect profiles of risperidone and clozapine in 20 out patients with schizophrenia. Am J Psych 1996; 153: 417–9Google Scholar
- 198.Zoleptil: summary of product characteristics. Newbury; Orion Pharma (UK) Ltd, 2002Google Scholar
- 199.Atypical antipsychotic drugs and stroke: message from Professor Gordon Duff, Chairman, Committee on Safety of Medicines (CEM/CMO/200401) [online]. Available from URL: http://www.info.doh.gov.uk/doh/embroadcast.nsf/vwDiscussionAll/3D8Dbb48B26FF90280256E520045977A [Accessed 2007 Oct 1]