Overview
- The SMART Trial tested whether a prespecified multimodal algorithm could guide choice between sertraline and bupropion and found no significant difference on the trial’s primary endpoint for biomarker-concordant drug assignment.
- Participants with two positive biomarkers had a 71.4% response rate, those with one positive biomarker had a 65.4% response rate, and participants with two negative biomarkers had a 42.9% response rate.
- The biomarker set combined a functional MRI marker, measures of reward learning and sensitivity, cognitive control tests, clinical variables (depression severity and neuroticism), and employment status to generate drug-specific indications.
- Predictive performance varied by drug with cross-validated AUCs of 0.86 for bupropion and 0.66 for sertraline, and authors say limited sample size reduced power to detect moderate effects on the primary outcome.
- If replicated in larger and more diverse samples and adapted for real-world use without routine MRI, biomarker stratification could shorten the trial-and-error period for patients and help clinicians identify people unlikely to respond to standard antidepressants so they can try alternatives sooner.