Habitual diets rich in dark-green vegetables are associated with an increased response to ω-3 fatty acid supplementation in Americans of African ancestry

Abstract

Although substantial variation exists in individual responses to omega-3 (ω-3) (n-3) fatty acid supplementation, the causes for differences in response are largely unknown. Here we investigated the associations between the efficacy of ω-3 fatty acid supplementation and a broad range of nutritional and clinical factors collected during a double-blind, placebo-controlled trial in participants of African ancestry, randomly assigned to receive either 2 g eicosapentaenoic acid (EPA) + 1 g docosahexaenoic acid (n = 41) or corn/soybean oil placebo (n = 42) supplements for 6 wk. Food-frequency questionnaires were administered, and changes in erythrocyte lipids, lipoproteins, and monocyte 5-lipoxygenase-dependent metabolism were measured before and after supplementation. Mixed-mode linear regression modeling identified high (n = 28) and low (n = 13) ω-3 fatty acid response groups on the basis of changes in erythrocyte EPA abundance (P < 0.001). Compliance was equivalent (∼88%), whereas decreases in plasma triglycerides and VLDL particle sizes and reductions in stimulated monocyte leukotriene B4 production were larger in the high-response group. Although total diet quality scores were similar, the low-response group showed lower estimated 2005 Healthy Eating Index subscores for dark-green and orange vegetables and legumes (P = 0.01) and a lower intake of vegetables (P = 0.02), particularly dark-green vegetables (P = 0.002). Because the findings reported here are associative in nature, prospective studies are needed to determine if dietary dark-green vegetables or nutrients contained in these foods can enhance the efficacy of ω-3 fatty acid supplements. This trial was registered at clinicaltrials.gov as NCT00536185.

Conflict of interest statement

Author disclosures: A. O’Sullivan, P. Armstrong, G. U. Schuster, T. L. Pedersen, H. Allayee, C. B. Stephensen, and J. W. Newman, no conflicts of interest.

Figures

FIGURE 1

RBC EPA response phenotype identification in Americans of African ancestry who received ω-3 FA supplementation for 6 wk. (A) Finite mixture regression analysis identified 2 distinct response phenotypes in the ω-3 FA–supplemented group, one showing a significant correlation between postintervention change in RBC EPA (i.e., ω-3 High; n = 29; r2 = 0.59, P < 0.001) and one that did not (ω-3 Low; n = 12; r2 = 0.07, P = 0.4). The placebo control also showed no correlation with compliance (n = 42; r2 = 0.02, P = 0.4). (B) Correcting changes in RBC EPA for baseline status and the mg/kg-consumed dose reassigned 1 high-response member to the low-response group and showed that the magnitude of response was greater in participants with the lowest baseline RBC EPA concentrations in both groups (ω-3 High: n = 28; slope = −0.02, y-intercept = 0.0129, r2 = 0.52, P < 0.0001; ω-3 Low: n = 13; slope = −0.02, y-intercept = -0.0038, r2 = 0.52, P = 0.005) as well as in the placebo controls (n = 42; slope = −0.008, r2 = 0.17, P = 0.006). Regression results ± 95% CIs are shown. Placebo data and regressions with nonsignificant slopes are not shown. Pre, presupplementation; ω-3 High, ω-3 high responders; ω-3 Low, ω-3 low responders.

FIGURE 2

RBC EPA enrichment in African Americans receiving fish oil supplementation for 6 wk correlates with monocyte capacity to produce immunomodulatory metabolites. Changes in RBC EPA correlated with A23187-stimulated monocyte EPA-derived 5-HEPE increases in both responsive groups (ω-3 High: n = 28; slope = 11, r2 = 0.55, P < 0.0001; ω-3 Low: n = 11; slope = 34, r2 = 0.57, P = 0.007) (A) and with arachidonic-derived LTB4 decreases in the high-response group only (ω-3 High: n = 29; slope = −2.0, r2 = 0.25, P = 0.005; ω-3 Low: n = 12; P = 0.98) (B). Placebo controls showed no changes in either variable. Regression results ± 95% CIs are shown. Placebo data and regressions with nonsignificant slopes are not shown. LTB4, leukotriene B4; 5-HEPE, 5-hydroxyeicosapentaenoic acid; ω-3 High, ω-3 high responders; ω-3 Low, ω-3 low responders.

FIGURE 3

Assessment of diet quality in Americans of African ancestry who received placebo or ω-3 FA supplements for 6 wk using modified HEI–2005 scoring. (A) The HEI categories are shown for the study groups and for the U.S. 2005 national average in non-Hispanic blacks. Although the study cohort differed significantly from the national population in multiple categories, calories from SOFAAs are not directly comparable (see text). The HEI–DGOL subscores without common letters differed by ANOVA (P < 0.05). (B) Study HEI-DGOL subscores ranged from 0.22 to 5.0, with 62% of the ω-3 Low group and 64% of the ω-3 High group found in the second (1.4–2.6) and fourth (3.8–5.0) quartiles, respectively. (C) The HEI-Total score did not differ between groups (ω-3 Low: 56 ± 4; ω-3 High: 64 ± 2; Placebo: 61 ± 1; P = 0.2) and was comparable to that in the 2005 non-Hispanic black population (HEI = 54). The ω-3 High group HEI-Total scores were skewed to the observed range (36–78) upper quartiles. DGOL, dark-green and orange vegetable plus legume; HEI, Healthy Eating Index; Non-Hisp, non-Hispanic; Q, quartile; SOFAA, solid fats, added sugars, and alcohols; Veg, vegetables; ω-3 High, ω-3 high responders; ω-3 Low, ω-3 low responders.