Text version
Let’s look at the use of antipsychotics in the treatment of acute schizophrenia. Are some antipsychotics better than others? Can they be rank-ordered by efficacy and tolerability? This is similar to the old CATIE trial from the US National Institute of Mental Health. The motive then was determining whether expensive second-generation antipsychotics were better. Here, the motivation is that we have many antipsychotics to choose from. Are any of them better than the others? Thanks to a research team that examined 54,000 publications, selecting 400 for study, we have a new perspective.
This one is another network meta-analysis similar to Quick Take, Volume 1, study number 4, which was from the same team. That study was led by Andrea Cipriani, this one by Maximilian Huhn, each in the Lancet. This technique aggregates data on effect size for medications across multiple studies, including comparisons against one another, as well as against placebo. Moreover, the comparisons examined here were not just about efficacy but also positive symptoms, negative symptoms, social functioning, and the big tolerability problems like weight gain, akathisia, sedation, among others.
They looked at all of the second-generation antipsychotics licensed in either Europe or the US, plus 17 first-generation agents, for a total of 32 different medications. The first problem with interpreting their data is that we surely do not use all 32 of these medications, and some are not licensed everywhere. We regularly use half of them at most. It is easy to look at the ones that we regularly use, ignore the rest, and see if we can detect any differences between our routine medications. Overall, the results somewhat mimic what we clinicians have generally figured out for the acute phase of schizophrenia, meaning, acute psychosis. The efficacy differences are quite minimal. Maybe clozapine sticks out a bit, and maybe amisulpride, but US psychiatrists do not have access to that one. According to this network meta-analysis, amisulpride causes nearly as much prolactin elevation as risperidone, which led the list for that problem along with its 9-hydroxy metabolite, paliperidone. Zotepine is high on the efficacy side but worse for weight gain, worse than clozapine and olanzapine. Notably, quetiapine and risperidone were not too far behind.
So, US psychiatrists are wondering if they should bemoan their lack of access to sertindole, which was the lowest on the list for all-cause discontinuation. However, it is not licensed by the FDA due to sudden deaths in the trials, and in this network meta-analysis, sertindole was associated with by far the most significant QT prolongation.
Back to the commonly used agents: Olanzapine causes more weight gain than quetiapine and risperidone, which were worse than aripiprazole and lurasidone. Ziprasidone was the same as placebo. For akathisia, risperidone and haloperidol were at the top of the list of commonly used agents. Lurasidone was right up there with them, with aripiprazole notably further down but not innocent.
There were some interesting statistical maneuvers in this paper coping with the study population differences, like age, gender, severity of illness, sometimes even using different outcome measures. They used a technique called CINeMA (confidence in network meta-analysis), which was developed before the study to characterize confidence in their findings, but no significant conclusions hinge on this statistical maneuver. Overall, I think the data support my sense that clinicians who are regularly using an array of medications will, over a short time, quite accurately appreciate differences in efficacy and side effects. For example, it did not take us very long to figure out that olanzapine caused far less extrapyramidal symptoms than risperidone but had a much higher propensity to cause weight gain, or that its efficacy for psychosis was not better than haloperidol. That is what we are getting confirmed in this network meta-analysis.
Abstract
Comparative Efficacy and Tolerability of 32 Oral Antipsychotics for the Acute Treatment of Adults With Multi-Episode Schizophrenia: A Systematic Review and Network Meta-Analysis
Maximilian Huhn, Adriani Nikolakopoulou, Johannes Schneider-Thoma, Marc Krause, Myrto Samara, Natalie Peter, Thomas Arndt, Lio Bäckers, Philipp Rothe, Andrea Cipriani, John Davis, Georgia Salanti, Stefan Leucht
Background: Schizophrenia is one of the most common, burdensome, and costly psychiatric disorders in adults worldwide. Antipsychotic drugs are its treatment of choice, but there is controversy about which agent should be used. We aimed to compare and rank antipsychotics by quantifying information from randomised controlled trials.
Methods: We did a network meta-analysis of placebo-controlled and head-to-head randomised controlled trials and compared 32 antipsychotics. We searched Embase, MEDLINE, PsycINFO, PubMed, BIOSIS, Cochrane Central Register of Controlled Trials (CENTRAL), WHO International Clinical Trials Registry Platform, and ClinicalTrials.gov from database inception to Jan 8, 2019. Two authors independently selected studies and extracted data. We included randomised controlled trials in adults with acute symptoms of schizophrenia or related disorders. We excluded studies in patients with treatment resistance, first episode, predominant negative or depressive symptoms, concomitant medical illnesses, and relapse-prevention studies. Our primary outcome was change in overall symptoms measured with standardised rating scales. We also extracted data for eight efficacy and eight safety outcomes. Differences in the findings of the studies were explored in metaregressions and sensitivity analyses. Effect size measures were standardised mean differences, mean differences, or risk ratios with 95% credible intervals (CrIs). Confidence in the evidence was assessed using CINeMA (Confidence in Network Meta-Analysis). The study protocol is registered with PROSPERO, number CRD42014014919.
Findings: We identified 54 417 citations and included 402 studies with data for 53 463 participants. Effect size estimates suggested all antipsychotics reduced overall symptoms more than placebo (although not statistically significant for six drugs), with standardised mean differences ranging from -0·89 (95% CrI -1·08 to -0·71) for clozapine to -0·03 (-0·59 to 0·52) for levomepromazine (40 815 participants). Standardised mean differences compared with placebo for reduction of positive symptoms (31 179 participants) varied from -0·69 (95% CrI -0·86 to -0·52) for amisulpride to -0·17 (-0·31 to -0·04) for brexpiprazole, for negative symptoms (32 015 participants) from -0·62 (-0·84 to -0·39; clozapine) to -0·10 (-0·45 to 0·25; flupentixol), for depressive symptoms (19 683 participants) from -0·90 (-1·36 to -0·44; sulpiride) to 0·04 (-0·39 to 0·47; flupentixol). Risk ratios compared with placebo for all-cause discontinuation (42 672 participants) ranged from 0·52 (0·12 to 0·95; clopenthixol) to 1·15 (0·36 to 1·47; pimozide), for sedation (30 770 participants) from 0·92 (0·17 to 2·03; pimozide) to 10·20 (4·72 to 29·41; zuclopenthixol), for use of antiparkinson medication (24 911 participants) from 0·46 (0·19 to 0·88; clozapine) to 6·14 (4·81 to 6·55; pimozide). Mean differences compared to placebo for weight gain (28 317 participants) ranged from -0·16 kg (-0·73 to 0·40; ziprasidone) to 3·21 kg (2·10 to 4·31; zotepine), for prolactin elevation (21 569 participants) from -77·05 ng/mL (-120·23 to -33·54; clozapine) to 48·51 ng/mL (43·52 to 53·51; paliperidone) and for QTc prolongation (15 467 participants) from -2·21 ms (-4·54 to 0·15; lurasidone) to 23·90 ms (20·56 to 27·33; sertindole). Conclusions for the primary outcome did not substantially change after adjusting for possible effect moderators or in sensitivity analyses (eg, when excluding placebo-controlled studies). The confidence in evidence was often low or very low.
Interpretation: There are some efficacy differences between antipsychotics, but most of them are gradual rather than discrete. Differences in side-effects are more marked. These findings will aid clinicians in balancing risks versus benefits of those drugs available in their countries. They should consider the importance of each outcome, the patients’ medical problems, and preferences.
Funding: German Ministry of Education and Research and National Institute for Health Research.
Download PDF and other files
Reference
Huhn, M., Nikolakopoulou, A., Schneider-Thoma, J., Krause, M., Samara, M., Peter, N., Arndt, T., Bäckers, L., Rothe, P., Cipriani, A., Davis, J., Salanti, G., & Leucht, S. (2019). Comparative efficacy and tolerability of 32 oral antipsychotics for the acute treatment of adults with multi-episode schizophrenia: A systematic review and network meta-analysis. The Lancet, 394(10202), 939-951.
