Genetics
Common vitamin and supplement use does not shift Lp(a) over time, UK Biobank analysis of 12,109 (Lipids Health Dis 2026)
Original title: Baseline self-reported supplement category use and two-visit difference in lipoprotein(a): an observational analysis in UK Biobank
UK Biobank observational analysis of 12,109 participants with Lp(a) measured at baseline and a repeat visit a median 4.4 years later, testing whether self-reported use of common supplement categories (vitamins A, B, C, D, E, folate, multi-vitamins) is associated with within-person change in Lp(a). Adjusted estimates for the log-scale two-visit difference were small and centred near null, ranging from -0.014 to 0.005 log10 units (roughly -3.18% to 1.22% relative change), and none met the Bonferroni-corrected significance threshold; results were unchanged with expanded clinical and genetic-risk adjustment. The authors conclude common supplement use is not associated with meaningful change in Lp(a), though only broad self-reported categories, not doses, were assessed.
Original abstract
Background: Circulating Lp(a) is largely genetically determined, but repeated measurements indicate that Lp(a) concentrations may differ within individuals over time. It remains unclear whether self-reported use of common supplement categories is associated with differences between repeated Lp(a) measurements. We aimed to investigate associations between baseline self-reported supplement-category use and two-visit difference in Lp(a), and to explore whether these associations differed by Lp(a) genetic risk.
Methods: We included 12,109 UK Biobank participants with Lp(a) measured at baseline and at the first repeat assessment. Self-reported use of common supplement categories (vitamins A, B, C, D, E, folate, and multi-vitamins) was assessed by questionnaire. The primary outcome was the continuous two-visit difference in log10[Lp(a)], defined as log10[repeat-assessment Lp(a)] minus log10[baseline Lp(a)]. The primary analysis used core behavioral/sociodemographic-adjusted multivariable linear regression. Threshold-category and genetic-risk analyses were secondary and exploratory.
Results: Among 12,109 participants with repeated Lp(a) measurements over a median 4.4 years, estimates in the primary core behavioral/sociodemographic-adjusted continuous log-scale model were small and centered near the null, ranging from β = -0.014 to 0.005 log10 units, corresponding to approximately - 3.18% to 1.22% relative difference in the conditional repeat-to-baseline Lp(a) ratio. None met the Bonferroni-corrected threshold. Expanded clinical/genetic adjustment and secondary descriptive screens did not alter the null interpretation.
Conclusions: In this selected repeat-assessment sample, baseline self-reported use of common supplement categories was not associated with two-visit difference in log10[Lp(a)]. Because supplement use was measured only as broad baseline self-reported categories, these findings should not be interpreted as evidence about dose-specific, sustained, or biological effects of vitamin supplementation on Lp(a).
Summary written by lp-a.org from the published abstract; figures as published. Page updated 17 August 2026. Methods.