Glyco-metabolic deteriorations are the most limiting adverse reactions to antipsychotics in the long term. They have been incompletely investigated and the properties of antipsychotics that determine their magnitude are not clarified.

To rank antipsychotics by the magnitude of glyco-metabolic alterations and to associate it to their pharmacological and chemical properties, we conducted a network meta-analysis.

We searched PubMed, Embase, and Psycinfo on 10 September 2020. We selected studies containing the endpoint-baseline difference or the distinct values of at least one outcome among glucose, HbA1c, insulin, HOMA-IR, triglycerides, total/HDL/LDL cholesterols. Of 2094 articles, 46 were included in network meta-analysis. Study quality was assessed by the RoB 2 and ROBINS-I tools. Mean differences (MD) were obtained by random-effects network meta-analysis; relations between MD and antipsychotic properties were analyzed by linear regressions. Antipsychotic properties investigated were acidic and basic pKa, polar surface area, polarizability, and occupancies of D2, H1, M1, M3, α1A, α2A, 5-HT1A, 5-HT2A, 5-HT2C receptors.

We meta-analyzed 46 studies (11 464 patients); on average, studies lasted 15.47 weeks, patients had between 17.68 and 61.06 years of mean age and 61.64% were males. Olanzapine and clozapine associated with greater deteriorations, aripiprazole and ziprasidone with smaller deteriorations. Higher polarizability and 5-HT1A receptor occupancy were associated with smaller deteriorations, H1, M1, and M3 receptor occupancies with larger deteriorations.

Drug rankings may guide antipsychotic switching toward metabolically safer drugs. Mechanistic insights may suggest improvements for combination therapies and drug development. More data are required regarding newer antipsychotics.