The Question

Obesity is a major public health problem. In those with severe mental illness, the prevalence of obesity is twice that of the general population. Obesity, hypertension, dyslipidaemia, and smoking contribute to the three fold increase in mortality observed in severe mental illness. This translates in to a 15 year reduction in life expectancy in this group. Antipsychotic medications are one of the established causes of weight gain. So clinicians and patients always look for weight neutral treatment options. This is a very difficult process as one need to find an effective antipsychotic that is weight neutral in an individual patient. Switching antipsychotic medications involves extremely difficult decision making as the risk of relapse and consequent harm to the individual can be substantial and the person may need resource intensive care to bring back stability. We all try to avoid switching from an otherwise effective medication. Life style interventions are often difficult in this context. Adding Metformin or glucagon-like peptide 1 receptor agonists (eg, exenatide, liraglutide) are other strategies used to combat weight gain. The two key questions here are 1. Is switching effective in reducing weight gain? 2. What is the cost appropriate agent for this purpose? Dan Siskind et al reports the results of a systematic review and meta-analysis in Schizophrenia Bulletin. (Advance publication/online).

The review included randomized controlled trials or observational studies of people with SMI that investigated a switch from any single or combination of antipsychotics to a different antipsychotic monotherapy for ≥4 weeks.Meta-analyses were carried out if at least 2 studies had useable data for the outcome of interest.


61 articles were included in the systematic review; 59 articles contributed to the meta-analyses.Mean study duration was 26.3 ± 19.7 weeks. Most studies included participants with schizophrenia. Study quality in general was graded as low. Lack of blinding in the RCTs, small sample sizes and high drop out rates were the common quality issues.

Two types of analysis were used. Pair wise analysis ( switch vs stay) made comparisons of switching to a specific antipsychotic vs remaining on the previous agent and Before- after meta-analysis summarised ‘before-to-after’ data of switching to a different antipsychotic.


  1. Switching to aripiprazole was associated with a 5.52 kg weight loss in the pairwise meta analysis. Before to after metanalysis showed a mean weight loss of 1.96 kg.
  2. Compared to staying on previous antipsychotics, switching to olanzapine was associated with a 2.46 kg weight gain ( pair wise meta analysis). Before after meta analysis showed that switching to olanzapine resulted in a 2.7 kg weight gain.
  3. There was no difference in weight after switching to amisulpride. Switch to lurasidone, risperidone/ paliperidone or quetiapine showed no weight change.Patients switched to clozapine gained 2.8 kg.
  4. Switching to ziprasidone showed a significant weight loss 
  5. Effect on psychotic symptoms: In pairwise analysis, there was no significant difference between switching to aripiprazole or staying on previous antipsychotic for psychotic symptoms. Similar was the case with olanzapine. In before- after metaanalsyis, switching to aripiprazole, paliperidone/risperidone, quetiapine, or ziprasidone were associated with a reduction in psychotic symptoms, while switching to olanzapine did not change psychotic symptoms. 


Aripiprazole had the best evidence for weight loss upon antipsychotic switching; 5.5 kg in the pairwise and 2.0 kg in the before-to-after analysis. Switching to ziprasidone was also associated with weight loss but based on only before-to-after data.Switching to amisulpride, paliperidone/risperidone, quetiapine, or lurasidone had no effect on weight.We need more good quality research to make better informed decisions.


Pairwise switch-vs-stay meta-analyses was possible only for aripiprazole and olanzapine. For the other antipsychotics, there were only usable data for before-to-after analyses. Before after analysis has limitations and it can overstate the effects. Quality of studies included are low and hence caution required on conclusions.

Practice implications:

Ziprasidone is not licensed in the UK. Ziprasidone’s potential cardiac effects is a reason for this. Drug reaction with eosinophilia (DRESS) is another rare but serious side effect to worry.

Safety regarding relapse/ psychotic symptoms in relation to switch requires further clarity.

The study:

Schizophrenia Bulletin, sbaa191,

Source link