Omega-3 fatty acid plasma levels before and after supplementation: Correlations with mood and clinical outcomes in the omega-3 and therapy studies

Objective: To examine fatty acid profiles, their response to omega-3 fatty acid (Ω3) supplementation, and associations with clinical status and treatment response in youth with mood disorders. Methods: In a placebo-controlled 2X2 design, 7-14 year-olds (N = 95) in parallel pilot trials (depression N = 72; bipolar N = 23) were randomly assigned to 12 weeks of Ω3 supplementation (1.4 g eicosapentaenoic acid [EPA], 0.2 g docosahexaenoic acid [DHA], and 0.27 g other Ω3 per day); psychoeducational psychotherapy (PEP); their combination; or placebo (mainly oleic and linoleic acid) alone. Blood was drawn at baseline (N = 90) and endpoint (n = 65). Fatty acid levels were expressed as percent of total plasma fatty acids. Correlational and moderator/mediator analyses were done with SPSS Statistics 23. Results: At baseline: (1) DHA correlated negatively with alpha-linolenic acid (ALA) (r = -0.23, p = 0.029); (2) Arachidonic acid (AA, Ω6) correlated negatively with global functioning (r = -0.24, p = 0.022); (3) Total Ω3 correlated negatively with age (r = -0.22, p = 0.036) and diastolic blood pressure (r = -0.31, p = 0.006). Moderation: Baseline ALA moderated response to Ω3 supplementation: ALA levels above the sample mean (lower DHA) predicted significantly better placebo-controlled response (p = 0.04). Supplementation effects: Compared to placebo, 2 g Ω3 per day increased EPA blood levels sevenfold and DHA levels by half (both p < 0.001). Body weight correlated inversely with increased EPA (r = -0.52, p = 0.004) and DHA (r = -0.54, p = 0.003) and positively with clinical mood response. Mediation: EPA increase baseline-to-endpoint mediated placebo-controlled global function and depression improvement: The greater the EPA increase, the less the placebo-controlled Ω3 improvement. Conclusion: Ω3 supplementation at 2 g/day increases blood levels substantially, more so in smaller children. A possible U-shaped response curve should be explored.