Weight and body mass index increase in children and adolescents exposed to antipsychotic drugs in non-interventional settings: a meta-analysis and meta-regression.

Antipsychotics increase weight, BMI and waist size, particularly in pediatric patients. Switching antipsychotics is common practice, thus defining the risk for each antipsychotic in real-life settings can be important for clinical guidance. We conducted a meta-analysis on antipsychotic-related changes in body measures in pediatric observational studies. Of 934 publications found on PubMed, we analyzed 38, including nine treatment arms: no treatment, mixed antipsychotic treatment, first-generation antipsychotics, aripiprazole, clozapine, olanzapine, quetiapine, risperidone and ziprasidone. Changes in weight, BMI, BMI-Z and waist size were meta-analyzed according to the duration of clinical observations: 6, 12, > 12 months. Meta-regressions probed influencing factors. Weight in Kg was increased at 6, 12, > 12 months by olanzapine [+ 10.91, + 10.7, data not available (n/a)], mixed antipsychotic treatment (n/a, + 9.42, + 12.59), quetiapine (+ 5.84, n/a, n/a) and risperidone (+ 4.47, + 6.01, + 9.51) and without treatment (n/a, + 2.3, n/a). BMI was increased at 6, 12, > 12 months by olanzapine (+ 3.47, + 3.42, n/a), clozapine (n/a, + 3, n/a) mixed antipsychotic treatment (+ 3.37, + 2.95, + 3.32), risperidone (+ 2, + 2.13, + 2.16), quetiapine (+ 1.5, + 1.82, n/a), aripiprazole (n/a, + 1.7, + 2.1) and without treatment (n/a, + 0.75, n/a). BMI-Z was increased at 6, 12, > 12 months by olanzapine (+ 0.94, + 0.98, + 0.89), clozapine (n/a, + 0.8, n/a), risperidone (+ 0.62, + 0.61, + 0.48), quetiapine (+ 0.57, + 0.54, n/a), mixed antipsychotic treatment (+ 0.51, + 0.94, + 0.44), without treatment (n/a, + 0.37, n/a) and aripiprazole (no gain, + 0.31, n/a). Waist size in cm was increased at 6, 12 months by risperidone (+ 8.8, + 11.5), mixed antipsychotics treatment (+ 9.1, + 10.2) and quetiapine (+ 6.9, + 9.1). Overall, olanzapine and clozapine displayed maximum risk, followed by risperidone, quetiapine and aripiprazole (more risky at longer terms); ziprasidone was associated with no gains. No time-based trends emerged, suggesting a drug-specific risk magnitude. Meta-regressions evidenced variable roles for persistence in therapy and follow-up length, increased risk for drug-naïve patients, and a ceiling effect determined by higher baseline BMI/BMI-Z values.