50. PHARMACOGENETICS TESTING TO OPTIMISE PRESCRIBING IN PSYCHOSIS

Background Several antipsychotics, and most antidepressants, have pharmacogenetic clinical guidelines as well as FDA/MHRA/EMA approved labels providing recommendations on drug selection or dose adjustments based on genetic data. However, implementation of pharmacogenetics testing in mental health practice remains limited. It is also unclear who should be offered such tests and if they are cost-effective for the healthcare system. Aims 1) To set up a testing pipeline in collaboration with people with lived experience and clinicians. 2) Investigate the influence of genetic variation on adherence to medication. 3) Study the clinical and cost effectiveness of pharmacogenetics in psychosis. 4) Evaluate acceptability to clinicians and patients. Methods Participants: Adults with schizophrenia or bipolar disorder of any age, sex or ancestry.Intervention: Prospective study offering pharmacogenetics testing to inform the choice and dose adjustment of antipsychotic and other psychotropic medicines. Pragmatic, single-blind trial that does not restrict patient/clinician choices. Comparator: Before/after pharmacogenetic intervention. Control group receiving treatment as usual with antipsychotic drugs. Outcomes: Adverse drug reactions, adherence, quality of life, cost-effectiveness.We tested a multi-gene panel selected from evidence-based pharmacogenetic clinical guidelines (CYP2D6, CYP2C19, CYP2B6, CYP3A4 and other genes). We also used genotyping arrays, PCR, and sequencing. We measured turnaround times and cost-effectiveness of different technologies. Linear regression was used to model genetic effects on adherence and adverse drug reactions to psychotropic drugs. Results Participants: 363 adults with psychosis. 46% females. Average age 43 ±15 years. This is a diverse sample where 13% of participants identified as Asian/British-Asian, 12% as Black/Black-British/Caribbean/African, 66% European/White and 9% as having multiple/other ancestries. Using a 5-gene panel, we found that 97% of patients carried one or more actionable variants for the prescribing of antipsychotic and/or antidepressant agents.At baseline, patients with a higher CYP2D6 activity score reported higher levels of adherence to medication as measured by the Medication Adherence Report Scale (MARS coef.=0.27; 95%CI: -0.05 to 0.60; p= 0.098). Patients with higher CYP2D6 activity scores also reported lower levels of adverse drug reactions measured by the Liverpool University Neuroleptic Side Effects Rating Scale (LUNSERS coef.= -0.03; 95% CI: -0.06 to -0.01; p= 0.017). Mediation analyses suggest that the CYP2D6 variation effects on adherence to medication are mediated by the frequency/severity of adverse drug reactions (ACME coef.=0.18; 95%CI: 0.02 to 0.35; p=0.032), rather than a direct genetic effect (ADE coef.= 0.10; 95%CI: -0.21 to 0.41; p=0.57). Discussion We have developed a new pipeline of pharmacogenetic testing and medication optimisation for people with psychosis (the first in UK mental health services, to the best of our knowledge). Our diverse sample is representative of the local population. Multi-gene panels offer the most cost-effective approach to testing. The observed influence of CYP2D6 variation on medication adherence is mediated by adverse drug reactions.