The choice of antipsychotic drug for the treatment of schizophrenia is controversial, and conventional pairwise meta-analyses cannot provide a hierarchy based on the randomised evidence. Researchers thus conducted a Bayesian-framework, multiple-treatments meta-analysis (which uses both direct and indirect comparisons) of RCTs to integrate the available evidence to create hierarchies of the comparative efficacy, risk of all-cause discontinuation, and major side-effects of antipsychotic drugs.
They analysed the results of RCTs in acute schizophrenia, comparing 15 antipsychotics, including all recently approved agents, and placebo across 212 trials with 43,049 participants. The primary outcome was efficacy, as measured by mean overall change in symptoms. They also examined all-cause discontinuation, weight gain, extrapyramidal side-effects, prolactin increase, QTc prolongation, and sedation.
In the analysis, all antipsychotics were significantly more effective than placebo, however, the efficacy differences were generally small. The standardised mean differences with 95% credible intervals were:
• clozapine 0.88, 0.73—1.03
• amisulpride 0.66, 0.3—0.78
• olanzapine 0.59, 0.53—0.65
• risperidone 0.56, 0.50—0.63
• paliperidone 0.50, 0.39—0.60
• zotepine 0.49, 0.31—0.66
• haloperidol 0.45, 0.39—0.51
• quetiapine 0.44, 0.35—0.52
• aripiprazole 0.43, 0.34—0.52
• sertindole 0.39, 0.26—0.52
• ziprasidone 0.39, 0.30—0.49
• chlorpromazine 0.38, 0.23—0.54
• asenapine 0.38, 0.25—0.51
• lurasidone 0.33, 0.21—0.45
• iloperidone 0.33, 0.22—0.43
Conversely, as has also been noted previously, differences in adverse effects were larger and sometimes in the opposite direction, such that more effective drugs also tended to have more adverse effects. The odds ratios compared with placebo for all-cause discontinuation ranged from 0.43 for the best drug (amisulpride) to 0.80 for the worst drug (haloperidol); for extrapyramidal side-effects 0.30 (clozapine) to 4.76 (haloperidol); and for sedation 1.42 (amisulpride) to 8.82 (clozapine). Standardised mean differences compared with placebo for weight gain varied from −0.09 for the best drug (haloperidol) to −0.74 for the worst drug (olanzapine), for prolactin increase 0.22 (aripiprazole) to −1.30 (paliperidone), and for QTc prolongation 0.10 (lurasidone) to −0.90 (sertindole).
The researchers note that the findings of their primary outcome challenge the dogma that the efficacy of all antipsychotic drugs is the same- a notion that originated from an influential narrative review published in 1969, which has not been scientifically addressed since. They suggest that this multiple-treatments meta-analysis provides evidence-based hierarchies for the efficacy and tolerability of antipsychotic drugs, overcoming the major limitation of conventional pairwise meta-analyses.
They also note several limitations to the study: the network could have been expanded to old drugs such as perphenazine and sulpiride, which have had good results in effectiveness studies, but only a few relevant perphenazine trials have been done. Furthermore, as more second-generation antipsychotics lose their patent protection, the debate about the costs of the original second-generation antipsychotics becomes less important, with the current debate really focusing on whether the newest drugs are cost effective. In addition, the findings cannot be generalised to young people with schizophrenia, patients with predominant negative symptoms, refractory patients, or stable patients, as they were all excluded to enhance homogeneity as required by multiple-treatments meta-analysis.
Overall, the researchers hope that the suggested hierarchies in the abovementioned seven major domains will help clinicians to adapt choice of antipsychotic drug to the needs of individual patients, and lead to modification of clinical practice guidelines.
An accompanying comment article acknowledges this as an exemplary piece of research but suggests the findings should be interpreted with caution, bearing in mind their limitations, to inform clinical decision making. However the authors agree with the investigators that the findings support the idea that each antipsychotic drug needs to be assessed on the basis of its individual risk-to-benefit profile and that generalising classifications into first- and second-generation antipsychotics is mostly not useful. They also highlight that data from studies, and even more so from meta-analyses, are based on group means, thus it is important for clinicians to remember that individuals can respond rather differently.
The NICE clinical guideline on schizophrenia states that the benefits and side-effect profile of each antipsychotic drug should be discussed with the patient and choice of therapy should be made by the patient and healthcare professional together, considering the relative potential of such treatment to cause extrapyramidal side effects (including akathisia), metabolic side effects (including weight gain) and other side effects (including unpleasant subjective experiences).