Meta-analysis Ranks Antipsychotics on Efficacy, Side Effects
July 18, 2013. With dozens of different antipsychotics on the market today, finding the best one for each individual patient is difficult. A new meta-analysis may make the choice a little easier by providing clinicians with rankings of 15 different antipsychotics across several categories. Published online June 27, 2013, in The Lancet, the study created hierarchies in terms of efficacy, several common side effects, and overall acceptability. Clozapine was the top-ranking antipsychotic in terms of efficacy, although there was no clear winner in terms of side effects.
“All drugs have their pros and cons, and these profiles should be used to adapt their use to the needs of individual patients,” said lead study author Stefan Leucht of Technische Universität München in Germany in an e-mail to SRF.
Due to the large number of antipsychotics that are available, many of them have not been compared head to head. Conventional meta-analyses that compare pairs of drugs are not able to create hierarchies that incorporate data for which no direct comparison has been made. For example, suppose drug A has been compared to drug B in a clinical trial, and drug B has been compared to drug C in a different trial. Without a direct comparison of drug A and C, a traditional meta-analysis can’t provide information on how the two compare. However, by using a multiple-treatment meta-analysis, Leucht and colleagues were able to circumvent this problem. The multiple-treatment approach allows both direct and indirect comparisons of antipsychotics—it can provide an indirect estimate of the comparison between A and C using the comparisons between A to B and B to C.
In an accompanying commentary, Christopher Correll of the Zucker Hillside Hospital in Glen Oaks, New York, and Marc De Hert of Belgium’s Katholieke Universiteit Leuven praise the methodology that Leucht and colleagues employed, noting that it “represents a laudable example of bringing to bear the strengths of the approach, while attempting to minimize its weaknesses.”
The meta-analysis included a total of 212 randomized, controlled trials, reported between 1955 and 2012, with 43,049 enrolled participants. Trials were collected from previous systematic reviews, several databases including the Cochrane Central Register of Clinical Trials, and unpublished data from pharmaceutical companies. Included studies were at least single-blinded (though over 90 percent of studies were double-blinded), and all trials examined antipsychotics taken orally as monotherapies. Participants had received a diagnosis of schizophrenia or schizoaffective, schizophreniform, or delusional disorder.
The primary outcome measure was change in symptoms as measured by the Positive and Negative Symptom Scale (PANSS). When PANSS data were not available, the Brief Psychiatric Rating Scale (BPRS) was used. Because a multiple-treatment meta-analysis requires reasonable homogeneity in the sample, trials done in pediatric patients or those with predominantly negative symptoms, concurrent medical illness, treatment resistance, or stable illness were excluded. For the same reason, all included studies measured acute treatment, which the authors defined as six weeks' duration (or the timepoint between four and 12 weeks that was closest to six weeks when six-week data were not available).
All 15 drugs were more effective than placebo (effect sizes ranged from -0.33 to -0.88). Consistent with earlier studies, the second-generation antipsychotic clozapine was significantly more effective than all other drugs and emerged as the “clear winner,” said Leucht. However, this finding is based largely on comparisons with first-generation drugs, he added, and thus future head-to-head trials that pit clozapine against other second-generation drugs are warranted. Amisulpride, olanzapine, and risperidone were also more effective than the rest of the drugs, except for paliperidone and zotepine. The authors noted the small but robust efficacy differences they observed “challenge the dogma that the efficacy of all antipsychotics is the same.”
The scoop on side effects
Overall, the side effect differences among drugs were much larger than the changes observed for efficacy, and there was no best or worst drug across all of the side effect domains. All drugs except haloperidol, ziprasidone, and lurasidone caused weight gain. Olanzapine was the worst offender, followed by zotepine.
Over half of the drugs were favorable compared to placebo in terms of extrapyramidal side effects (EPS), movement disorders classically associated with first-generation antipsychotics. Clozapine produced significantly fewer EPS than all other drugs, followed by sertindole, olanzapine, and quetiapine. Haloperidol was again at the bottom of the list, producing more EPS than all drugs except for zotepine and chloropromazine. In addition, only four drugs (amisulpride, paliperidone, sertindole, and iloperidone) were not more sedating than placebo.
The authors also examined the effect of the antipsychotics on prolactin concentration. High levels of this hormone, a condition termed hyperprolactinaemia, have been associated with sexual dysfunction, an absence of menstrual periods, inappropriate lactation, and infertility (Haddad and Wieck, 2004). Only five drugs (aripiprazole, quetiapine, asenapine, chloropromazine, and iloperidone) did not elevate prolactin levels compared to placebo, while paliperidone and risperidone ranked worst, producing a greater increase in prolactin concentrations than the 11 other drugs for which data were available.
Lurasidone, aripiprazole, paliperidone, and asenapine were the only four antipsychotics not associated with prolongation of the QTc interval, a measure of the heart’s electrical cycle that, when elongated, can lead to dangerous cardiac complications.
The results of the new study suggest that several generalizations about the differences between older, first-generation (typical) and newer, second-generation (atypical) antipsychotics are oversimplifications. For example, although first-generation drugs are typically thought to be the main culprits of EPS, several second-generation antipsychotics also produced significantly more EPS than did placebo. In addition, second-generation antipsychotics are generally blamed for weight gain and metabolic side effects, though the first-generation drug chloropromazine was one of the worst offenders in this category.
When the researchers examined all-cause discontinuation, a global measure that incorporates both efficacy and side effect tolerance to index drug acceptability, they found that all drugs, with the exception of zotepine, were better than placebo. Five drugs (amisulpride, olanzapine, clozapine, paliperidone, and risperidone) had the lowest all-cause discontinuation rates, indicating better acceptability, than the others. Haloperidol was the least acceptable.
A new tool for ranking
Based on the data, Leucht and colleagues created individual hierarchies for efficacy, each of the five side effects measured, and all-cause discontinuation. The hope is that clinicians will be able to use these drug profiles to choose antipsychotics accordingly to a patient’s specific needs. For example, a patient with known cardiac problems could be placed on one of the drugs that is unassociated with a prolonged QTc interval.
Although results from multiple-treatment meta-analyses “must be cautiously interpreted,” the fact that the current one replicated many previous findings raises his confidence in the method, Leucht told SRF. However, meta-analyses are, by definition, based on group means, note Correll and De Hert. “That individuals can respond rather differently is important for clinicians to remember,” they concluded.—Allison A. Curley.
Correll CU, Hert MD. Antipsychotics for acute schizophrenia: making choices. Lancet. 2013 Jun 26. Abstract
Leucht S, Cipriani A, Spineli L, Mavridis D, Orey D, Richter F, Samara M, Barbui C, Engel RR, Geddes JR, Kissling W, Stapf MP, Lässig B, Salanti G, Davis JM. Comparative efficacy and tolerability of 15 antipsychotic drugs in schizophrenia: a multiple-treatments meta-analysis. Lancet . 2013 Jun 26. Abstract