Bayes A, Parker G. How to choose an antidepressant medication.Objective: We consider how to choose an antidepressant (AD) medication for the treatment of clinical depression. Method: A narrative review was undertaken addressing antidepressant 'choice' considering a range of parameters either weighted by patients and clinicians or suggested in the scientific literature. Findings were synthesised and incorporated with clinical experience into a model to assist AD choice. Results: Efficacy studies comparing ADs offer indicative guidance, while precision psychiatry prediction based on genetics, developmental trauma, neuroimaging, behavioural and cognitive biomarkers, currently has limited clinical utility. Our model offers guidance for AD choice by assessing first for the presence of a depressive subtype or symptom cluster and matching choice of AD class accordingly. Failing this, an AD can be chosen based on depression severity. Within-class choice can be determined by reference to personality style, patient preference, medical or psychiatric comorbidities and side-effect profile. Conclusion: Clarification of AD choice would occur if medications are trialled in specific depressive subtypes rather than using the generic diagnosis of major depressive disorder (MDD). Such 'top-down' methods could be enhanced by 'bottom-up' studies to classify individuals according to symptom clusters and biomarkers with AD efficacy tested in these categories. Both methods could be utilised for personalised AD choice.• Our model offers guidance of AD choice by assessing for a depressive subtype (e.g. melancholic), symptom cluster or weighting by severity and matching choice of AD class accordingly.• Within-class AD choice can be determined by reference to factors such as personality style, patient preference, medical or psychiatric comorbidities.• Clarification of AD choice would occur if medications are trialled in those with specific depressive subtypes rather than in relation to MDD, with complementary 'bottom-up' studies classifying individuals according to symptom clusters and biomarkers with AD efficacy tested in these categories.
Additional Comments• The dominant use of the major depressive disorder (MDD) diagnosis homogenises a variable set of depressive conditions which obscures the ability to detect treatment effects and inform antidepressant choice.• Extrapolating AD trial efficacy data to 'real-world' clinical practice is limited by trials excluding those with more severe depressive disorders.• There is a lack of agreement in validly defining depressive 'subtypes', as well as the relationship between surface symptoms and underlying endophenotypes.