Real-world omega-3 supplementation patterns and whole-blood fatty-acid biomarkers: a cross-sectional analysis of 762,150 dried blood spot samples

Abstract

Abstract Background Long-chain omega-3 polyunsaturated fatty acids, including eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), as well as compositional measures such as the arachidonic acid-to-EPA ratio (AA/EPA), have been extensively examined in fatty-acid biomarker research and in relation to cardiometabolic risk. However, large-scale characterisation of how these biomarkers and broader fatty-acid profiles vary across real-world omega-3 supplementation-exposure categories remains limited. This study aimed to characterise whole-blood fatty-acid profiles, with particular emphasis on EPA + DHA, AA/EPA, and their distributions across three pre-existing real-world supplementation-exposure groups in a large direct-to-consumer DBS dataset. Methods A retrospective cross-sectional analysis was performed using 762,150 whole-blood DBS samples submitted to a commercial fatty-acid profiling service between 2012 and 2023. Based on self-reported supplementation exposure, samples were classified as non-user, self-supplemented (≥ 120 days), or program-supplemented (≥ 120 days within a structured program including a standardized fish-oil/extra-virgin olive-oil formulation, weight-based dosing guidance, and biomarker testing/feedback). Whole-blood fatty-acid profiles, EPA + DHA, and ln(AA/EPA) were compared across groups after adjustment for age, sex, BMI, and geographic region. Additional analyses examined AA/EPA across comparable EPA + DHA intervals and regional variation in AA/EPA. Results Supplementation group was associated with differences in the measured relative whole-blood fatty-acid profile, with the largest group-related effect sizes observed for long-chain n-3-related measures and AA/EPA. In primary adjusted models, supplementation group had partial η² values of 0.274 for EPA + DHA and 0.352 for ln(AA/EPA). Adjusted mean EPA + DHA was progressively higher across the non-user, self-supplemented, and program-supplemented groups (3.30%, 4.16%, and 5.61%, respectively), while adjusted geometric mean AA/EPA was progressively lower (16.6, 10.9, and 5.2, respectively). The self-supplemented group showed intermediate but broadly dispersed biomarker values. By contrast, the program-supplemented group showed a more compact AA/EPA distribution, lower AA/EPA at comparable EPA + DHA levels below approximately 10%, and less regional heterogeneity in AA/EPA. Conclusions In this large cross-sectional DBS dataset, pre-existing real-world supplementation-exposure groups were associated with differences in covariate-adjusted EPA + DHA and AA/EPA, and biomarker dispersion. Simple self-reported supplement-use categories captured group-level biomarker differences but masked substantial within-category heterogeneity. AA/EPA provided complementary information on fatty-acid profiling not summarized by measured EPA + DHA alone and may help characterize biomarker patterning across real-world omega-3 supplementation contexts. These observational findings should be interpreted as associations rather than evidence of causal effects of specific supplementation strategies.

The authors' abstract, as published at the source. Lipids in Health and Disease, 2026 · DOI ↗