Background Previous meta-analyses have shown that almost all antipsychotics are associated with weight gain. However, mean weight gain is not informative about clinically relevant weight gain or weight loss. Aims To provide further insight into the more severe body weight changes associated with antipsychotic use, we assessed the proportion of patients with clinically relevant weight gain (CRWG) and clinically relevant weight loss (CRWL), defined as ≥7% weight gain and ≥7% weight loss. Method We searched PubMed, Embase and PsycInfo for randomised controlled trials of antipsychotics that reported CRWG and CRWL in study populations aged 15 years or older. We conducted meta-analyses stratified by antipsychotic and study duration using a random-effects model. We performed meta-regression analyses to assess antipsychotic-naive status and psychiatric diagnosis as modifiers for CRWG. PROSPERO: CRD42020204734. Results We included 202 articles (201 studies). Almost all included antipsychotics were associated with CRWG. For CRWL, available data were too limited to draw firm conclusions. For some antipsychotics, CRWG was more pronounced in individuals who were antipsychotic-naive than in individuals switching to another antipsychotic. Moreover, a longer duration of antipsychotic use was associated with more CRWG, but not CRWL. For some antipsychotics, CRWG was higher in people diagnosed with schizophrenia, but this was inconsistent. Conclusions Switching antipsychotic medication is associated with both weight gain and weight loss, but the level of CRWG is higher than CRWL in antipsychotic-switch studies. CRWG was more pronounced in antipsychotic-naive patients, highlighting their vulnerability to weight gain. The impact of diagnosis on CRWG remains inconclusive.

Background In recent years, antipsychotic-induced weight gain (AIWG) has gained more attention in research. Although interventions to prevent weight gain are currently being investigated, AIWG remains a major problem for both patients and clinicians and often results in poor treatment adherence, a decrease in quality of life. Furthermore, schizophrenia is associated with higher mortality rates and a decreased life expectancy. Recently, some new antipsychotic drugs have been introduced that are hypothesised to entail no or low incidences of clinically relevant weight gain (CRWG), and high incidences of clinically relevant weight loss (CRWL). Here ‘clinically relevant’ is defined as >7% weight change. In this meta-analysis, we aim to give a complete overview of both CRWG and CRWL, including these newer antipsychotics. Methods We searched Pubmed, Embase and Psychinfo for randomized clinical trials of antipsychotics that reported 7% weight change in study populations aged 18 years or older. We performed meta-analyses stratified by study duration (<6 weeks, 6–16 weeks, 16–38 weeks and >38 weeks) with a random effects model. Results The search yielded in total 941 articles. Ninety-two articles could be included in the meta-analysis, resulting in 341 records in the data set. All data were related to AP switch patients, no data on AP-naïve patients were found. During SIRS final results will be presented. Preliminary results showed that haloperidol, paliperidone and quetiapine had relatively low CRWG (16.6%, 18.7% and 18.4%, respectively), aripiprazole and risperidone had relatively high percentages of CRWG (25.4% and 24.0%, respectively). Olanzapine (29.5%) and lurasidone (7.4%) resulted in respectively the highest and lowest CRWG at >38 weeks of treatment. In the placebo group, CRWG was 3.8%. Incidences of CRWG continued to rise even after 38 weeks of treatment in most treatment groups. CRWL occurred with all antipsychotic drugs; at 6–16 weeks aripiprazole (7.9%) and ziprasidone (7.1%) had CRWL similar to placebo (8.7%). We found insufficient data on CRWL in the long term (>38 weeks) to draw any conclusions. Discussion All antipsychotics can result in both weight gain and weight loss. Previous research showed that patients more often gain weight than lose weight (Bak, 2014) and this is replicated in the present meta-analysis. Proportions CRWG and CRWL seem different between the antipsychotics. Future network meta-analysis are needed to test statistical significance of those differences. It appears, however, that CRWG is higher in patients receiving antipsychotics drugs compared to placebo. No conclusions can be drawn on CRWL due to insufficient data. It is clear that after >38 weeks of treatment, no ‘plateau’ phase is reached as CRWG continued to increase. More future research is needed on long-term weight effects on both CRWG and CRWL to give a clear overview on the ‘real’ effects on weight, as the majority of studies had a duration of less than 26 weeks. Furthermore, more research is needed on the long-term dose-response relationship in CRWG, as this could prove to be a method for managing weight in some patients.