Send to

Choose Destination
World Psychiatry. 2018 Oct;17(3):330-340. doi: 10.1002/wps.20579.

Tardive dyskinesia risk with first- and second-generation antipsychotics in comparative randomized controlled trials: a meta-analysis.

Carbon M1, Kane JM1,2,3,4, Leucht S4, Correll CU1,2,3,5.

Author information

Department of Psychiatry, Zucker Hillside Hospital, Glen Oaks, NY, USA.
Department of Psychiatry and Molecular Medicine, Hofstra Northwell School of Medicine, Hempstead, NY, USA.
Center for Psychiatric Neuroscience, Feinstein Institute for Medical Research, Manhasset, NY, USA.
Department of Psychiatry and Psychotherapy, Klinikum rechts der Isar der Technischen Universität München, Munich, Germany.
Campus Virchow-Klinikum, Charité-Universitätsmedizin Berlin, and Department of Child and Adolescent Psychiatry, Berlin Institute of Health, Berlin, Germany.


Tardive dyskinesia (TD) risk with D2/serotonin receptor antagonists or D2 receptor partial agonists (second-generation antipsychotics, SGAs) is considered significantly lower than with D2 antagonists (first-generation antipsychotics, FGAs). As some reports questioned this notion, we meta-analyzed randomized controlled studies (RCTs) to estimate the risk ratio (RR) and annualized rate ratio (RaR) of TD comparing SGAs vs. FGAs and SGAs vs. SGAs. Additionally, we calculated raw and annualized pooled TD rates for each antipsychotic. Data from 57 head-to-head RCTs, including 32 FGA and 86 SGA arms, were meta-analyzed, yielding 32 FGA-SGA pairs and 35 SGA-SGA pairs. The annualized TD incidence across FGA arms was 6.5% (95% CI: 5.3-7.8%) vs. 2.6% (95% CI: 2.0-3.1%) across SGA arms. TD risk and annualized rates were lower with SGAs vs. FGAs (RR=0.47, 95% CI: 0.39-0.57, p<0.0001, k=28; RaR=0.35, 95% CI: 0.28-0.45, p<0.0001, number-needed-to-treat, NNT=20). Meta-regression showed no FGA dose effect on FGA-SGA comparisons (Z=-1.03, p=0.30). FGA-SGA TD RaRs differed by SGA comparator (Q=21.8, df=7, p=0.003), with a significant advantage of olanzapine and aripiprazole over other non-clozapine SGAs in exploratory pairwise comparisons. SGA-SGA comparisons confirmed the olanzapine advantage vs. non-clozapine SGAs (RaR=0.66, 95% CI: 0.49-0.88, p=0.006, k=17, NNT=100). This meta-analysis confirms a clinically meaningfully lower TD risk with SGAs vs. FGAs, which is not driven by high dose FGA comparators, and documents significant differences with respect to this risk between individual SGAs.


Tardive dyskinesia; annualized incidence; aripiprazole; clozapine; first-generation antipsychotics; meta-analysis; randomized controlled studies; schizophrenia; second-generation antipsychotics

Supplemental Content

Full text links

Icon for Wiley Icon for PubMed Central
Loading ...
Support Center