1. Sedation of acutely agitated patients: Liquid formulations and rapidly dissolving tablets are useful to safeguard drug intake (C). Parenteral administration is often only needed in emergency situations when oral medications are not feasible (C). A well studied intervention is i.m. haloperidol combined with i.m. promethazine (A). Combining i.m. haloperidol with i.m. lorazepam or i.m. promethazine is more effective than haloperidol alone (A). Newer antipsychotics have advantages in terms of extrapyramidal side-effects (A). Overall, many options are available without clear superiority of one in all situations. In emergency situations, pragmatic issues (availability, ease of administration) as well as familiarity are critical considerations, and clinicians may wish to develop local protocols based on the evidence above that is most relevant to their situations (C).
There are many strategies for the treatment of agitated patients, such as monotherapy with an antipsychotic (high or low-potency, oral or parenteral , i.e., short-acting injectable or intravenous), monotherapy with a benzodiazepine, combining both, adding a sedating low-potency FGA, rapidly dissolving tablets, liquid and inhalable formulations etc. Liquid and rapidly dissolving tablets are options for safeguarding drug intake. Parenteral administrations are not more efficacious per se although they may have a more rapid onset of action, and are usually only necessary in emergency situations when patients refuse medication. The combination of haloperidol and lorazepam/promethazine seems to be more efficacious than monotherapy (Alexander et al., 2004; Battaglia et al., 1997; Huf et al., 2007). The newer antipsychotics aripiprazole, olanzapine and ziprasidone are available as intramuscular formulations (Citrome, 2007a). Their advantage includes fewer EPS than haloperidol. A general problem is that the studies on these newer agents were conducted for registration and all participants had to give informed-consent. These are not the patients for whom short-acting intramuscular medication is indicated (often violent patients who are not willing to take medication). From a methodological perspective, four large (200-300 participants), real-world, pragmatic trials with intramuscular medications need to be highlighted (Alexander et al., 2004; Huf et al., 2007; Huf et al., 2003; Raveendran et al., 2007). In these, the combination of haloperidol and promethazine produced more rapid tranquillization with fewer dystonic reactions than haloperidol alone (Huf et al., 2007) and no more EPS than lorazepam alone (Alexander et al., 2004). It did beat olanzapine in secondary outcomes (number of additional injections) with no difference in EPS, probably due to the anticholinergic properties of promethazine (Raveendran et al., 2007). Only intramuscular midazolam, a highly sedating benzodiazepine was more efficacious than haloperidol plus promethazine (Huf et al., 2003), but the mean dose of haloperidol in this study was lower compared to the other listed trials and so a dosing-effect leading to the favorable outcome for the benzodiazepine compound cannot be completely ruled out. In summary, a definite recommendation of a single strategy is not possible. In emergency situations, pragmatic criteria (availability, ease of administration) as well as familiarity are critical considerations and clinicians may wish to develop local protocols based on the evidence above that are most relevant to their situations.
2. Predominant persistent negative symptoms: Antipsychotics reduce general negative symptoms, but the evidence for negative symptoms that persist after acute treatment is limited (A). Rule out secondary negative symptoms due to extrapyramidal side-effects (e.g., by lowering the dose, a trial of anticholinergic medication, or switching to an antipsychotic with a low EPS risk) (C). Avoid secondary negative symptoms by using a drug with a low EPS risk (C). Low-dose amisulpride (50-300mg/day) is the best studied drug for this indication and is the only one that has been shown to be more efficacious than placebo in several trials (A). Adding an antidepressant can be effective (A-).
“General negative symptoms”: Many acutely ill patients with schizophrenia present with both positive and negative symptoms. In such patients, all antipsychotics that have been evaluated have been shown to improve negative symptoms more than placebo and some new generation drugs (amisulpride, clozapine, olanzapine, risperidone) have been found to be more effective than older antipsychotics (mostly haloperidol and chlorpromazine) (Leucht et al., 2009a).
Almost all relevant studies were conducted in patients suffering predominantly from positive symptoms and the doses of haloperidol were higher than those now recommended. In such patients, it is unclear whether the small advantage for negative symptoms related to primary negative symptoms or only to secondary negative symptoms, which can stem from positive symptoms, depression or extrapyramidal side-effects. Secondary negative symptoms due to EPS may be ruled out by lowering the antipsychotic dose, switching to an antipsychotic with a lower EPS risk or by a trial with anticholinergic medication. Therefore, such negative symptoms may resolve by the application of an antipsychotic alone and we would not recommend starting with a combination strategy.
“Predominant persistent negative symptoms”: Once the acute phase with positive symptoms has resolved, patients often present with persistent negative symptoms.
Only studies in patients with primary/predominant negative symptoms, whose positive symptoms are low and remain stable, can clarify whether a medication is effective for this indication. Another major problem is that negative symptoms and symptoms of depression overlap and are difficult to disentangle by rating scales, and that several antipsychotic drugs also have antidepressant effects (Komossa et al. 2009). Nevertheless, the best evidence is for low-dose amisulpride (50-300mg/day) which proved superior compared to placebo in several trials (Leucht et al., 2002), but even this finding is complicated by a single trial in which only olanzapine 5mg/day, but not amisulpride 150mg/day or olanzapine 20mg/day were more effective then placebo. Only a few RCTs comparing antipsychotic drugs head-to-head in predominant negative symptoms are available (Leucht et al., 2009c), and among these only a single small trial showed a superiority of olanzapine compared to haloperidol (Lindenmayer et al., 2007). Several meta-analyses showed that adding an antidepressant may be effective for predominant negative symptoms (Rummel et al., 2006) or in patients with chronic schizophrenia (Sepehry et al., 2007; Smith et al., 2001), or in any patients with schizophrenia and negative symptoms (Singh et al., 2010). Again, to disentangle symptoms of depression and negative symptoms may have been difficult. For example, For example, in one trial, exclusion of depressed patients led to disappearance of the previously noted significant improvement in negative symptoms with adjunctive antidepressant use (Saha et al., 2007).
3. Depressive symptoms: Depressive symptoms associated with an acute episode of schizophrenia should not be automatically treated with an antidepressant, because they may resolve with the treatment with an antipsychotic alone (A). If there is a suspicion of antipsychotic induced depressive symptoms, these should be ruled out by dose reduction, anticholinergic medication or by switching to a drug with a lower EPS risk (C). Persistent or post-psychotic depression may be treated with an antidepressant added to the antipsychotic (A). The risk to provoke an exacerbation by adding an antidepressant is low (A-).
Depressive symptoms are frequently present in acutely ill patients with schizophrenia (Hausmann and Fleischhacker, 2002). Because they are associated with positive symptoms and may thus improve with antipsychotics alone (Leucht et al., 2009a), we would not recommend adding an antidepressant immediately. Moreover, in a meta-analysis some newer antipsychotic drugs (amisulpride, aripiprazole, clozapine, olanzapine, quetiapine) had superior antidepressant properties compared to older drugs (mainly haloperidol and chlorpromazine) in people with schizophrenia (Leucht et al., 2009b). Some newer antipsychotics have also obtained an officially approved indication for major depressive disorder in some countries (e.g. quetiapine in the EU and the US, aripiprazole in the US; or amisulpride for dysthymia in Italy). Moreover, older literature suggested that drugs which produce a lot of EPS (in particular first generation antipsychotics, such as haloperidol or fluphenazine) may produce “akinetic depression” (Van Putten, 1978), i.e., depression-like symptoms. We would therefore recommend switching to a drug with low EPS risk in such occasions.
Studies that demonstrated the effectiveness of antidepressants for depression systematically excluded people diagnosed with schizophrenia. While it is reasonable to assume that depressive symptoms in the context of schizophrenia might also improve with antidepressant treatment, scant evidence supports this. A Cochrane review found some evidence that antidepressants are effective for post-psychotic depression, but the number of trials was small and most (but not all) used tricyclic antidepressants, which are currently not first choice (Whitehead et al., 2003). The risk that an adjunctive antidepressant may provoke psychosis has been judged to be small (Siris, 1993; Whitehead et al., 2003).
4. Manic symptoms: Manic symptoms associated with an acute episode should not be automatically treated with an anti-manic treatment because they may resolve by treatment with an antipsychotic alone (C). If manic symptoms persist, the addition of an anti-manic drug, such as lithium, valproate or carbamazepine, may be tried (C).
It should be noted that the recommendation and the following text refers to manic symptoms associated with schizophrenia, and not to schizoaffective disorder where different rules may apply. Manic symptoms can be present in patients with schizophrenia. They may often remit when treated with an antipsychotic alone, because many antipsychotics have been shown to be effective for bipolar mania (Cipriani et al., 2011). We would therefore not generally recommend combining an antipsychotic with an anti-manic drug in the presence of manic symptoms. However, if manic symptoms cannot be controlled with an antipsychotic alone, the addition of lithium, valproate or carbamazepine may be warranted, although meta-analyses have not found a superiority of adding these drugs to antipsychotics for manic symptoms. These meta-analyses were, however, limited by the fact that most of the included studies focused on positive symptoms so that very few data on manic symptoms were available (Basan and Leucht, 2004; Leucht et al., 2007b; Leucht et al., 2007c). Single randomized trials showed that adding carbamazepine (Klein et al., 1984), valproate (Casey et al., 2003) (not replicable in a later study (Casey et al., 2009) or lithium (Biederman et al., 1979) to antipsychotics reduces mania/”excitement” associated with schizophrenia.
5. Obsessive-compulsive symptoms: if such phenomena are symptoms of the psychosis, the antipsychotic treatment must be optimized (C). An SSRI might be helpful (C). If obsessive-compulsive symptoms are thought to be side-effects of the antipsychotic drug, switching the antipsychotic is an option (C).
The dilemma is that obsessive compulsive symptoms can be part of the schizophrenia symptomatology. But there are also case reports suggesting that obsessive-compulsive symptoms are side-effects of antipsychotic drugs (Scheltema Beduin et al., 2012; Schirmbeck and Zink, 2012) . This makes theoretical sense, as many of the newer antipsychotics have the opposite mechanism of action as SSRIs in that they block serotonergic receptors (Kim et al., 2012). Many of these case reports were on clozapine or risperidone, but this does not necessarily mean that the phenomenon is limited to these drugs (Scheltema Beduin et al., 2012). To disentangle both factors can be clinically very difficult and often impossible. One criterion is whether the obsessive compulsive symptoms were present before antipsychotic treatment or whether they started only afterwards. The guideline group is not aware of RCTs for patients with both schizophrenia and obsessive compulsive symptoms. Pragmatically, if obsessive compulsive symptoms are part of the psychosis, the antipsychotic treatment must be optimized. An SSRI might also be tried, but should be stopped if ineffective. If obsessive compulsive symptoms are thought to be side-effects of the antipsychotic drug, switching the antipsychotic, perhaps to one without a prominent serotonergic blockade, may be considered.
6. Anxiety: short-term addition of a benzodiazepine can be useful (A-).
A systematic review found no convincing evidence for benzodiazepines in schizophrenia apart from rapid tranquillisation of acutely agitated patients (Dold et al., 2012)). A few old RCTs (Azima et al., 1962; Gundlach et al., 1966; Minervini et al., 1990; Morphy, 1986) included in this Cochrane review examined the effects of benzodiazepines in people with schizophrenia and anxiety, and some mentioned positive effects. These studies were, however, typically small and often so poorly reported that they could not be used for meta-analytic calculations. The guideline group, nevertheless, felt that due to the well-documented anxiolytic effects of benzodiazepines in non-psychotic patients, adding a benzodiazepine to treat anxiety symptoms may be useful. Similar logic suggests that a SSRI might be tried, but the group is not aware of any RCT on this strategy and would also not generally recommend their use in this indication based on the personal experiences of the guideline group.
7. Insomnia: Attempt first non-pharmacological approaches, such as sleep-hygiene advice (C); if insufficient alone, apply together with pharmacologic options (C). Antipsychotic drugs with a sedating profile can be useful (C). If this is not sufficient, benzodiazepines/non-benzodiazepine hypnotics, sedating antipsychotics or sedating antidepressants can be added (C).
Certain antipsychotics are more sedating than others (see Table 1 and (Leucht et al., 2013a)). Such a sedating property can be useful in patients suffering from insomnia, although we are not aware of RCTs in patients with schizophrenia selected for insomnia, which would be the rigorous test of this approach. We are also not aware of comparative randomized controlled trials on other options for patients with schizophrenia and insomnia, such as adding benzodiazepines/non-benzodiazepine hypnotics or sedating antidepressants, or adding to switching to sedating antipsychotics. Such drugs should be chosen according to their side-effect profiles, risk for drug-drug interactions with the antipsychotic, and potential for abuse. Furthermore, as in other patients, sedative-hypnotics should be tried after, and in concert with, non-pharmacological approaches and sleep-hygiene advice.
8. Cognitive dysfunction: Antipsychotics improve cognitive function compared to placebo but the benefits are small (A). There is little evidence to guide the choice of antipsychotics for this indication. Acetylcholinesterase inhibitors and memantine have generally not been shown to be efficacious (A).
In contrast to a long-held opinion that first-generation antipsychotic drugs worsen cognitive function, a meta-analysis found a small, but statistically significant improvement of such antipsychotics compared to placebo (Mishara and Goldberg, 2004). Advantages of any specific antipsychotic are not very clear. A meta-analysis found that, as a group, clozapine, olanzapine, quetiapine and risperidone improved cognitive function modestly more than haloperidol (Woodward et al., 2005). However, in CATIE, perphenazine was at least as effective as olanzapine, quetiapine, risperidone and ziprasidone for cognition (Keefe et al., 2007), and in EUFEST there was no difference between low-dose haloperidol and amisulpride, olanzapine, quetiapine and ziprasidone (Davidson et al., 2009). There is no evidence that antipsychotics improve cognitive deficits outside of the context of treating people with psychotic illnesses.
Trials of acetylcholinesterase inhibitors to improve cognitive deficits mostly have not proven to be effective (Stip et al., 2007) although a study with galantamine showed improvement on select neuropsychological functions compared to placebo (Buchanan et al., 2008). There is no convincing randomized evidence on memantine for cognition in schizophrenia. Although one of two very small trials (less than 30 participants showed a superiority of memantine compared to placebo added to antipsychotics (de Lucena et al., 2009), the other one found no difference (Lee et al., 2012), and the largest RCT was negative (Lieberman et al., 2009).
9. Chronic aggressive behavior: Address the root of aggression. For example, if aggression is associated with positive symptoms, optimized treatment of symptoms and drug adherence are key (C). Clozapine was the most effective drug in an appropriate trial in chronically aggressive patients (A). The evidence for other newer antipsychotics or anti-manic drugs, such as lithium, carbamazepine or valproate, is methodologically less convincing (C).
Citrome (Citrome, 2007b, c) stressed the importance of considering the root of aggression. If it is due to positive symptoms, optimized treatment of psychopathology and adherence are key (for the role of depot treatment see maintenance treatment below). If aggression is caused by drug or alcohol abuse, or by social problems, these need to be addressed. In terms of pharmacological interventions, the guideline group is only aware of one antipsychotic drug trial that was conducted in patients specifically selected for chronic aggressive behavior. In this study, clozapine was more effective than haloperidol and olanzapine to reduce aggression, and this effect was independent of improvements in psychopathology (Krakowski et al., 2006). The evidence for other (newer) antipsychotic drugs is less strong, because it was derived from post-hoc analyses of RCTs that were conducted in patients who were not necessarily aggressive. In EUFEST olanzapine had an advantage over haloperidol, amisulpride, quetiapine and ziprasidone in managing hostility (Volavka et al., 2011). Therefore, it was difficult to disentangle antiaggressive effects from effects on other symptoms, limiting generalizability . The same problem holds true for lithium and antiepileptic drugs, for which the evidence in schizophrenia is scant (Citrome, 2007b, d; Citrome et al., 2007)). The recommendation for other treatments, such as beta-blockers or SSRIs, was even more cautious (Citrome, 2007a, b).
10. Suicidality: clozapine has been shown to be efficacious in such patients (A).
The guideline group is only aware of one large randomized controlled trial in people with schizophrenia selected for suicidality. Clozapine was more effective than olanzapine in this study (Meltzer et al., 2003). In a post hoc analysis sertindole reduced suicidal attempts to a greater extent than risperidone (Crocq et al., 2010).
11. Dual diagnosis: Antipsychotics are appropriate treatments for schizophrenia when there is a co-occurring substance use disorder, being likely effective for schizophrenia symptoms (A-). There is no strong evidence to support the preferential use of any antipsychotic (C). A long-acting injectable antipsychotic may be helpful in this situation to improve medication adherence (C). Antipsychotics do not seem to exert an effect on primary alcohol or cocaine/psychostimulant dependence (A).
Few RCTs in dual diagnosis patients showed significant improvement in schizophrenia symptoms and mixed/unclear results for substance use (Akerele and Levin, 2007; Rubio et al., 2006; Sayers et al., 2005). However, studies were generally small and compared two antipsychotics head-to-head (without clear group differences) and lacked a placebo group, so that the effect of time cannot be ruled out. Of note, in two recent meta-analyses, there was no beneficial effect of antipsychotics for primary substance abuse in alcohol dependent (Kishi et al., 2013) and cocaine dependent substance abusers without schizophrenia (Amato et al., 2007). Approved medications for substance dependence can be tried, but evidence of benefit is scarce. One RCT showed that naltrexone and disulfiram are more effective than placebo in promoting abstinence from alcohol in people with psychotic disorders (Petrakis et al., 2006). There is no evidence from RCTs supporting the use of acamprosate for people with schizophrenia. Psychosocial substance abuse treatments should be offered in addition to pharmacologic interventions (Dixon et al. PORT recommendations (Dixon et al., 2010)).
12. Adolescents: Antipsychotic drugs are efficacious in children and adolescents with schizophrenia (A). Drug tolerability should guide the choice of antipsychotic (A). Clozapine has shown to be a more efficacious antipsychotic in adolescents with treatment resistant schizophrenia (A).
Based on successful, 6-week, placebo controlled trials, newer antipsychotics, such as aripiprazole, olanzapine, paliperidone, quetiapine and risperidone, have FDA indication for schizophrenia in patients ages 13-17, and in this age group only ziprasidone failed to separate from placebo (Schimmelmann et al., 2013). The few head to head comparisons have not shown differences in terms of different efficacy between antipsychotics in this age range (Arango et al., 2009; Kumra et al., 1996; McClellan et al., 2007; Sikich, 2008; Sikich et al., 2008), with the exception of more efficacy for clozapine in adolescents with treatment refractory schizophrenia (Kumra et al., 2008). Although no direct comparison with the adult population has been conducted, a number of side effects are more frequent in adolescents compared to adults. These include sedation, EPS, withdrawal dyskinesia, prolactin elevation, weight gain and lipid abnormalities (Correll et al., 2006). In addition, the developing brain and body may have more long lasting effects from drug side effects. In adolescents, side effects profile vary widely between different antipsychotics (Correll et al., 2009a; Fraguas et al., 2011). For instance, the numbers-needed-to-harm for weight gain ≥7% went from 39 (confidence interval [CI]: -1 to +6, not significant) with aripiprazole to 3 (CI: 3-4) for olanzapine (De Hert et al., 2011c). Based on the metabolic side effect profile, there is no rationale for using olanzapine as first line treatment in adolescents. Metabolic adverse events should be monitored with all antipsychotics early during treatment, as in adolescents and in never treated or first episode patients, cardiometabolic effects are most pronounced (Correll et al., 2009a).
13. First episode patients: Overall they respond better to antipsychotic drugs than chronic patients and seem to need lower doses (B) which could be targeted at the lower end of officially registered ranges (C). Side-effects may therefore play an even higher role in choosing among antipsychotic drugs than in more chronic patients (C), although some evidence suggests in part similar efficacy differences between drugs as in chronic populations (A).
Although the guideline group is not aware of a systematic review that compared drug efficacy in first-episode patients with more chronic patients, there are many examples in the literature suggesting that the former respond better to drugs. For example, in a pivotal study comparing risperidone with haloperidol in “general” schizophrenia patients, 63% responded to the best risperidone dose, but response was based on only at least 20% PANSS reduction from baseline (Peuskens and Group., 1995). In a first-episode study comparing risperidone and haloperidol, again 63% responded, but the response threshold was much higher (at least 50% PANSS reduction, (Emsley and Group, 1999)). Furthermore, in the large effectiveness study CATIE, overall, 11.7% of chronic patients achieved remission as defined by (Andreasen et al., 2005), (Levine et al., 2011), compared to 30% of first-episode patients in the effectiveness study EUFEST (Boter et al., 2009). In (Robinson et al., 1999b) and (Lieberman et al., 2003), more than 80% of first-episode patients achieved a remission of positive symptoms (different criteria than those by (Andreasen et al., 2005)) within one year. Supported by the fact that large scale pragmatic trials have shown clear advantages of new generation antipsychotics over even a low dose of haloperidol (eg. EUFEST; Kahn et al., 2008) especially also with respect to motor side effects (Rybakowski et al., 2014), the guideline group recommends that side-effect profiles should play an even greater role in drug choice for first-episode patients. While Crossley et al. 2010 found no difference between the groups of first-generation and second-generation antipsychotic drugs (Crossley et al., 2010), a more comprehensive meta-analysis only replicated the lack of differences between these two drug groups for total psychopathology and positive symptoms (Zhang et al., 2012). By contrast, second-generation antipsychotics were significantly better for negative symptoms, cognition and relapse prevention (small effect size) and were associated with significantly lower all-cause discontinuation rates and discontinuation due to inefficacy and intolerability (25-40% risk reduction), findings that seemed independent of haloperidol comparator doses (Zhang et al., 2012). However, the first-generation comparator was almost exclusively haloperidol and we feel that this broad classification is not valid because both first- and second-generation antipsychotics are heterogeneous groups (Leucht et al., 2013a). While the average doses in first-episode studies with flexible dosing were often relatively low (e.g. (Emsley and Group, 1999)), there are no evidence-based estimates. However, we would suggest to target doses at the lower end of the officially registered dose ranges.
14. Elderly patients with schizophrenia: the same principles apply, but this group is more sensitive to side-effects of antipsychotic drugs (B) and the target doses should be lower than for younger adult patients (C).
Due to various factors, such as decreased drug metabolism, lower plasma protein concentrations, elderly patients are more sensitive to antipsychotic side-effects. For example, the risk for tardive dyskinesia in elderly patients has been repeatedly documented to be considerably higher than in younger adults (Caligiuri et al., 2000). Therefore, side-effects play an even more important role for drug choice in this population. Similarly to first-episode patients, doses at the lower end of officially registered dose ranges are often sufficient and better tolerated.