Studies examining associations between antidepressant response and plasma levels of bupropion and its metabolites have yielded contradictory findings. There have been no such studies in youth. This study explored such associations in 8 boys and 8 girls, age 11 to 17 years, all prescribed bupropion sustained release (SR) for major depression (n = 6) or depressive disorder not otherwise specified (n = 10) as part of a pharmacokinetic (PK) study. All were started on morning doses of bupropion SR of 100 mg/day, and most eventually had doses increased to 200 mg/day because of inadequate clinical response. After taking prescribed dose of bupropion SR at least 14 days (median = 21 days), subjects had steady-state serial plasma levels of bupropion and its metabolites measured during a 24-hour period after morning doses. A total of 9 subjects underwent these PK assessments on doses of 100 mg/day, and 6 underwent these on doses of 200 mg/day, with 4 studied on both doses. In this 24-hour assessment, the treating psychiatrist rated subjects' antidepressant response using the Clinical Global Impression's Improvement scale (CGI-I), blind to plasma levels, but informed by child and parent rating scales of depressive symptoms and clinical interviews. Relative to 7 nonresponders, 9 responders (CGI-I < or = 2) had significantly higher mean areas under concentration curves for bupropion (P = 0.03), threohydrobupropion (P = 0.02), and erythrohydrobupropion (P = 0.02), and especially hyroxybupropion (P = 0.006). Plasma levels 7.5 hours after morning doses reaching the following cut points discriminated responders from nonresponders: bupropion > or = 37 ng/mL (P = 0.001), hydroxybupropion > or = 575 ng/mL (P = 0.003), threohydrobupropion > or = 240 ng/mL (P = 0.009), or erythrohydrobupropion > or = 45 ng/mL (P = 0.009). These preliminary findings suggest that plasma levels of bupropion and metabolites, particularly hydroxybupropion, may predict acute antidepressant response in depressed youths taking bupropion SR.