|Home | About | Journals | Submit | Contact Us | Français|
People with a first or recurrent psychotic episode tend to present late for medical attention, and many do not present at all. Presentation is often initiated by others, not by patients themselves. Psychosis can also become apparent during a manic presentation, when patients act on their delusions in a public forum, or when they have the complications of substance misuse. Patients who experience intolerable symptoms (distressing delusions or voices; box 1) often seek medical help. In emergency settings, family members' concerns contrast with the patient's apparent indifference. The highest risk of suicide in people with schizophrenia occurs during the first five years of illness (“the critical period”), and interventions are most fruitful during this time. Importantly, patients experiencing their first episode should quickly be given competent assessments and access to appropriate services.
I searched HighWire and PubMed from 2002 to October 2006 for randomised controlled trials and systematic reviews of treatments for psychosis and schizophrenia. I reviewed the Cochrane Database of Systematic Reviews using both terms, across the age spans. Influential articles, key texts, and published treatment guidelines are also included.
The one year prevalence of non-organic psychosis is 4.5 per 1000 community residents.1 Most new cases arise in men under 30 and women under 35, but a second peak occurs in people over 60 years. Psychotic symptoms had a 10.1% prevalence in a non-demented community population over 85 years.2 Schizophrenia has a one year prevalence of 3.3 per 1000 people, and a lifetime morbidity risk of 7.2 per 1000 people.w1 Independent of known associations with migration and ethnic origin, increased economic inequality in areas of high deprivation also predicts a higher incidence of schizophrenia.3 Some people who become depressed (one in five of us over a lifetime) also develop hallucinations and delusions, related to and “congruent with” their low mood. Bipolar affective disorder has a lifetime prevalence of 1.3-1.6%,4 and it is characterised by episodes of psychosis during both high (“manic”) and low (depressive) relapses. The misuse of substances, notably cannabis,5 raises the prevalence of psychotic symptoms further—substance misuse partly explains the 10 times higher prevalence of psychosis in prison populations.1 Psychosis occurs frequently in all forms of dementia including Parkinson's disease. Other causes of organic psychoses are neurological disorders (epilepsy, head injury, haemorrhage, infarction, infection, and tumours) and most causes of delirium. Taken together, therefore, acute psychosis is one of the most common psychiatric emergencies. There are explanations of psychotic “symptoms” other than the biomedical model of this review; medicalising psychosis as “an illness like any other” increases both public pessimism about outcome and the stigma attached to people with psychosis.6
Diagnosis is based on clinical findings. No confirmatory laboratory or radiological tests are available, although investigations are needed to rule out organic psychosis. More information is gained on first assessment than at any subsequent time: even a few days' antipsychotic treatment can reduce the strength of delusions, and patients learn quickly that disclosing too many symptoms can have implications for the drugs they are prescribed or their liberty.
Carefully defining the psychotic processes clarifies their nature. Patients' trust can be gained by recording presenting complaints first and listening empathically to their accounts of troubling symptoms. Open questions (“How have things been for you lately?”) should be followed by progressively more closed questions (“Do you think something funny has been going on? Have you heard unusual noises or voices? Could someone be behind this?”). Patients rarely complain of negative symptoms (box 2), but they may have lost ambitions at school or work and social networks and activities may have been curtailed. Three core mood symptoms—mood, energy, and interest or pleasure—are reduced in depression and raised in manic states. The coexistence of psychosis and major alterations in mood may indicate bipolar or schizoaffective disorders.4
Many other aspects of the patient's history determine diagnosis and management:
The primary diagnosis may be revised weeks or years later (box 3), and thorough documentation improves diagnostic accuracy now and later. The patient's general appearance and behaviour may indicate overarousal and hostility (as a result of positive symptoms) or irritability suggestive of elevated mood. Other motor signs (catatonia and negativism) are rare in Western settings. Altered consciousness is highly unusual in non-organic psychoses—intermittent clouding indicates delirium and this or other impairments require urgent medical investigation. Speech will be disorganised if thought disorder is present (box 1), and with predominant negative symptoms (box 2) conversation will be stilted and difficult. Random changes of the subject (loosening of associations) and new words (neologisms) are best written down verbatim. Fast or pressured speech suggests mania. Mood should be noted as normal, depressed, or elevated. Affect, the outward expression of mood, is unlikely to be normal in these patients: flat affect may be the most obvious sign of negative symptoms, but there may be others (box 2). An anxious or perplexed affect may impact on actual behaviour. Suicidality (thoughts, intentions, actions) must always be assessed by asking questions like, “Have the voices suggested suicide?” Other abnormalities of thought (obsessions, overvalued ideas) and perception (illusions, misinterpretations) are common. Cognitive impairment, tested at the bedside, can present in the early stages of psychosis, but gross abnormalities may alert the clinician to learning disability or organic pathology. Concentration is subjectively normal (patient unaware) but objectively impaired (for example, the patient cannot recite the months of the year backwards). Insight can change considerably over the course of a psychotic illness and its treatment.
Taking a collateral history is the third core component of assessments. Incomplete information can be gained from patients who are paranoid, and perhaps lack insight. Family and friends may have noticed them behaving strangely—responding to hallucinations or testing their delusions. Taking collateral details after clinical assessment is an opportunity to test the working diagnosis. The patient's family will clarify whether some beliefs are culturally sanctioned and are not therefore delusional. Collateral history may identify a prodrome or negative symptoms (box 2) as the main focus of carers' concerns. Insidious onset and prolonged psychotic symptoms during the first two years are both strong predictors of poorer long term outcome.7 If no prodrome has occurred and the episode has had a short duration (fewer than two weeks), with a clear stressful precipitant, the diagnosis might be one of an acute and transient psychotic disorder (box 3), which is best treated with support and without drugs. If the quality of collateral history is poor (for example, the patient is brought in by the police or is homeless), seek out anyone with prior contact before concluding your assessment. Integrate new information into further assessments of your patient: disclosure improves as a trusting relationship develops.
Investigations (table 11)) are led by positive physical findings that suggest organic causes or comorbidities. Patients need to be given the results as soon as they are known. This reduces anxiety and paranoia, and it prevents excessive preoccupation with physical health. If the patient has organic psychosis, treat the underlying condition and unless the patient has epilepsy8 consider symptomatic short term treatment of the psychotic symptoms. If the underlying condition cannot be cured (for example, Alzheimer's dementia), consider giving low dose antipsychotics, but be aware that the benefits of these drugs in this situation may be outweighed by their adverse effects.w2 An identified physical cause is best treated in a general hospital with support from psychiatrists. Even in agitated patients who lack insight into their bizarre behaviour, mental health law allows for compulsory treatment of the mental disorder only, and physical treatments (antibiotics, intravenous fluids, surgery) cannot be forced on patients under this legislation.
From this point on, I will consider only psychiatric diagnoses. The first three diagnoses listed in box 3 should be treated in the same way because they are all a “first episode.” Decisions to prescribe mood stabilisers are usually made during subsequent episodes of bipolar affective disorder and schizoaffective disorder.4 10 Of the remaining diagnoses (box 3), depression is the most common and easily treated,11 and like all affective psychoses has a relatively good prognosis.12 Delusional disorder is a persistent, non-organic, non-affective psychotic disorder, without protracted hallucinations or negative symptoms. Post-traumatic stress disorder and obsessive compulsive disorder have prominent anxiety symptoms, driven by understandable (non-psychotic) processes. The last three possibilities are lifelong disorders, identified by collateral history. All can have episodes of psychosis, but residual disabilities of the underlying disorder persist beyond the treatment episode.
I find that having psychosis is horrible, but unless I'm acting strangely no one knows and I'm expected to seem normal. I hear very distressing voices all the time and occasionally get weird delusions and see things in a way that other people say are not real. I've been admitted to hospital and sectioned several times because of it.
When I first arrive at the hospital I hate the fact that my liberty has been curtailed, but after a while it's a relief not to have the responsibility of trying to take care of myself. I know it's time to go home when I start resenting the hospital again.
Finding the right medication can be difficult—I have the misfortune of getting terrible side effects from many of them. However, by trial and error I have eventually found something that doesn't make me too uncomfortable and makes the voices quieter. Now of course I'm reluctant to try the new ones in case they cause problems or don't work properly.
The most complicated thing in day to day life is trying to work out what sensory input is real without having to keep asking people. I also have to try to make sure I don't get tired or stressed.
Non-organic psychoses are best treated by mental health services in the least restrictive setting. Even with sophisticated community services, more than 70% of patients with a first episode of psychosis are admitted to psychiatric hospital. Open discussion can achieve consensual admission. Patients with psychosis who decline further treatment are assessed under mental health legislation on the grounds of danger to self (suicide, unsafe behaviours, exploitation by others) or danger to others (overarousal, risk of acting on delusions, potential harm to others). In England and Wales, the Mental Health Act (1983) requires two independent doctors and an approved social worker to agree on involuntary committal to a psychiatric hospital. Psychosis is a disruptive and distressing experience; inpatient units need to be adapted to support rather than confront, and seclusion or “intensive care units” should be considered as last resorts. Accreditation for acute inpatient mental health services (www.rcpsych.ac.uk) sets out minimum and desirable standards. Older adults, adolescents, and postpartum women have complex needs and require admission to specialist units. Early detection, perhaps through specialised teams with allied strategies (public education, liaison with schools and general practitioners) have the potential to reduce admissions.13
One antipsychotic drug should be given at the lowest effective dose (table 22).). It is safer to achieve sedation with benzodiazepines (as required), rather than antipsychotics. Choosing a highly sedating antipsychotic drug at this stage can impede discharge later (box 4). All hospitals and trusts have clear guidelines on rapid tranquillisation of patients with psychosis; these monitor for side effects and complications.10 Typical antipsychotic drugs cause extrapyramidal side effects and raised prolactin (which causes sexual dysfunction and galactorrhoea) at therapeutic doses, as do most atypical antipsychotic drugs at higher doses. Typical antipsychotics have greater anticholinergic (dry mouth, tachycardia, urinary obstruction, etc) and antiadrenergic (postural hypotension, impotence) effects. All antipsychotics cause sedation to varying degrees and lower seizure threshold, especially clozapine. All antipsychotics, except for ziprasidone (not available in the UK) and aripiprazole,10 cause weight gain and impaired glucose tolerance. Typical and atypical antipsychotics probably increase the risk of thromboembolic disease equally.w3 Amisulpride and aripiprazole offer relatively lower risks of QTc prolongation on electrocardiography.10 Interactions between subclinical effects on the cardiovascular system and the metabolic syndrome are possible beyond the first episode, so body weight and blood monitoring (table 11)) should be repeated in three months, then yearly. The metabolic syndrome (box 5) is often seen in people with chronic psychoses given their unhealthy life styles,14 and is linked to all antipsychotics, most notably atypicals. The syndrome comprises a combination of insulin resistance and its physiological consequences—for example, a 20% increase in major coronary events over 10 years.15 Waist circumference (box 5) seems to be the best predictor of metabolic syndrome.
One meta-analysis reported that five of seven atypical drugs had significant advantages over typical antipsychotics in the treatment of acute psychosis; quetiapine and aripiprazole did not achieve significance (table 22).w4 The advantages of atypical drugs over typical drugs in the longer term are more marginal,w5 and one non-industry trial found no additional benefits of atypicals (excluding clozapine) for either patient preference or quality of life markers.w6 A multicentre trial of haloperidol and olanzepine measured brain volume in seven primary “magnetic resonance imaging regions of interest” over two years.9 Patients in the haloperidol group lost grey matter volume, mostly during the first 12 weeks of treatment, and these losses correlated with negative clinical outcomes. They could not conclude that haloperidol failed to prevent volume loss, or caused it, or that olanzepine halted loss.9 The choice of antipsychotic drug is best made in consultation with the patient, in the context of psychosocial interventions that promote recovery. Clozapine should be considered earlier in patients who do not respond to two antipsychotics, but many patients decline clozapine because weekly blood tests are needed to detect early signs of a low white blood cell count. Fish oils are a worthwhile option, but only as an addition to standard treatment.10
Given the adverse effects of antipsychotic drugs, non-pharmacological treatments (table 33)) should be more widely available, but this seems to be limited by ward culture and lack of suitably trained staff. Evidence supporting psychological interventions is strong enough to recommend their use in all treatment guidelines.17 Cognitive behaviour therapy reduces the impact of symptoms,w25 and family interventions prevent relapse.w26 Both achieve an agreed, individualised plan to recognise relapse earlier. Practitioners of cognitive behaviour therapy have challenged traditional assumptions about delusions to gain a shared understanding of the origins of beliefs and explore alternative explanations: disturbances in intrinsic thinking and subcultural beliefs (ideas common in the young adult age group most likely to develop non-organic psychosis) may underlie delusions.18 Family interventions are best carried out early in the course of illness, while the patient still lives at home. Family interventions have the advantage of benefits for other family members and greater acceptability and than drug treatments (drop-out rates are lower). Psychoeducational interventions are brief, cheap, and require less staff training. Most patients with a first episode of psychosis have misused substances; abstinence improves their prognosis, and if they continue to abstain their outcomes at 18 months are better than those for patients who have never misused substances.19
Box 6 shows possible outcomes of standard care, on the basis of two reviews.7 20 Between 30% and 60% of patients with a first episode of psychosis receiving interventions from UK community mental health teams had a good outcome at three years.12 Among those patients who do not respond to treatment initially, 16% have a good recovery at 15 years.7 These patients, who have a complex long term illness and a high risk of relapse, are best managed in specialist settings. Early intervention teams provide phase specific treatments, integrated case management, and cognitive behaviour therapy interventions.13 A randomised clinical trial has shown the value of integrated care in patients with a first episode of psychosis.21 Patterns of symptoms change over time and a modular form of cognitive behaviour therapy meets people's needs most effectively. Early intervention teams reduce the duration of untreated psychosis.13 22 Despite early misgivings among researchers, duration of untreated psychosis is a remediable, independent predictor of worse outcome.w35 Social functioning and vocational outcomes at 18 months are significantly improved by early intervention teams.23 Excess deaths including suicide are seen in all patients with a first episode of psychosis,7 but a fall in suicide rates in people with schizophrenia has been attributed to reduced access to lethal methods and better treatment (from early intervention teams).24 The evidence supporting early intervention teams is better than that justifying the current practice of standard care.w36 One key research question remains: “Do specialised early intervention teams offer improvements in outcome over and above those provided by phase-specific interventions alone?”w36
After recovery (full or partial), a single antipsychotic drug is given prophylactically, usually at a lower dose than that needed for treating acute illness (table 22).). Treatment of a first episode is recommended for one year, followed by gradual cessation in asymptomatic patients at low risk. Risk of relapse is indicated by residual disability, family history of psychosis, or current substance misuse. Patients at risk and those with multiple psychotic episodes require longer prophylaxis. Patients with a history of violence need more intensive case management to reduce risk, and this may include prolonged medication under supervision. Given the high personal and health service costs of relapse, decisions about discontinuation and prophylaxis should be agreed with early intervention teams. Several early models for intervention teams have been described, with varying resource implications.25 The subject of treatment resistant psychosis has been discussed by others.10 26 Multiple coordinated interventions at adequate doses with verified adherence, including clozapine as a third line drug, must be applied before “treatment failure” is confirmed.
Early intervention teams accommodate diagnostic uncertainty in some patients, and for most patients coordinated interventions maximise functioning and prevent relapse. These teams and their advocates (users, carers, and people outside the health professions) have brought enthusiasm and innovation to a large number of patients in whom “cure” was previously the exception rather than the rule. We still await the “perfect” antipsychotic drug, where improved efficacy is not undermined by disabling, initially hidden, side effects. Trials of cognitive behaviour therapy and family interventions, with more sophisticated treatments than used heretofore, will identify specific components that improve recovery and reduce relapse further.
Competing interests: None declared.
Provenance and peer review: Commissioned; peer reviewed.