Background: Antipsychotics are widely used and have major benefits in the treatment of psychiatric disorders. However, there are also well-documented risks including serious metabolic, cardiac, haematological, hepatic and pancreatic adverse reactions. It is not clear what the relative risks for these serious adverse effects are between the different antipsychotic drugs.

Aims: The aim of the present study is to compare the risks of serious adverse events (SAEs) between the antipsychotics used in clinical practice in New Zealand.

Methods: Exposure to antipsychotics was identified from the New Zealand National Pharmaceutical Collection. Outcomes were identified from three sources: the Suspected Medicine Adverse Reaction Search (SMARS) database of Medsafe, the National Coronial Information System (NCIS) and Hospital Separation Data. Data were linked using encrypted national health numbers. Data were analysed using descriptive statistics and incidence rates. All data were de-identified, the research used administrative databases of previously collected information and the research was given ethics approval at an institutional level.

Results: For the time period 2007 to 2015, there were 233,068 users and 4,528,875 non-users included in the analysis. There were 129,620 patients treated with quetiapine, 56,817 with risperidone, 38,897 with haloperidol, 30,767 with olanzapine, and 5,755 with clozapine. The incidence rate ratio (IRR) for arrhythmia was, for users compared to non-users, 3.26, and for haloperidol compared to non-users was 8.03. When analysed by age group, the IRR for all antipsychotics decreased with age from 15.36 in the 10 to 19 years age group to 0.84 in the over 80 years age group. Haloperidol had a relatively high risk for most of the outcomes. In the SMARS database, from 2007 to 2015, there were 3340 reports related to antipsychotics with 1292 (38.7%) related to clozapine. Clozapine was related to 86 (73.5%) reports of neutropenia, 56 (69.1%) of myocarditis and 42 (75.0%) of constipation. In the NCIS database, from 2008 to 2012 there were 31 deaths associated with clozapine, 24 with olanzapine, 21 with quetiapine and <5 with risperidone.

Conclusions: SAEs should be considered in the benefit risk assessment for each patient when choosing between alternative antipsychotics.