Effect of green-Mediterranean diet on intrahepatic fat: the DIRECT PLUS randomised controlled trial

Abstract

Objective: To examine the effectiveness of green-Mediterranean (MED) diet, further restricted in red/processed meat, and enriched with green plants and polyphenols on non-alcoholic fatty liver disease (NAFLD), reflected by intrahepatic fat (IHF) loss.

Design: For the DIRECT-PLUS 18-month randomized clinical trial, we assigned 294 participants with abdominal obesity/dyslipidaemia into healthy dietary guidelines (HDG), MED and green-MED weight-loss diet groups, all accompanied by physical activity. Both isocaloric MED groups consumed 28 g/day walnuts (+440 mg/day polyphenols provided). The green-MED group further consumed green tea (3-4 cups/day) and Mankai (a Wolffia globosa aquatic plant strain; 100 g/day frozen cubes) green shake (+1240 mg/day total polyphenols provided). IHF% 18-month changes were quantified continuously by proton magnetic resonance spectroscopy (MRS).

Results: Participants (age=51 years; 88% men; body mass index=31.3 kg/m²; median IHF%=6.6%; mean=10.2%; 62% with NAFLD) had 89.8% 18-month retention-rate, and 78% had eligible follow-up MRS. Overall, NAFLD prevalence declined to: 54.8% (HDG), 47.9% (MED) and 31.5% (green-MED), p=0.012 between groups. Despite similar moderate weight-loss in both MED groups, green-MED group achieved almost double IHF% loss (-38.9% proportionally), as compared with MED (-19.6% proportionally; p=0.035 weight loss adjusted) and HDG (-12.2% proportionally; p<0.001). After 18 months, both MED groups had significantly higher total plasma polyphenol levels versus HDG, with higher detection of Naringenin and 2-5-dihydroxybenzoic-acid in green-MED. Greater IHF% loss was independently associated with increased Mankai and walnuts intake, decreased red/processed meat consumption, improved serum folate and adipokines/lipids biomarkers, changes in microbiome composition (beta-diversity) and specific bacteria (p<0.05 for all).

Conclusion: The new suggested strategy of green-Mediterranean diet, amplified with green plant-based proteins/polyphenols as Mankai, green tea, and walnuts, and restricted in red/processed meat can double IHF loss than other healthy nutritional strategies and reduce NAFLD in half.

Trial registration number: NCT03020186.

Keywords: epidemiology; fatty liver; magnetic resonance imaging; nutrition.

Figure 1

Flow chart of the Dietary Intervention Randomized Controlled Trial Polyphenols Unprocessed study. HDG, healthy dietary guidelines; MED, Mediterranean.

Figure 2

(A–C)18-month changes in weight and intrahepatic fat. (A) 18-month absolute change in weight between intervention groups (ITT analysis, n=294). (B)18-month changes in IHF% between intervention groups (ITT analysis, adjusted p values for age, sex and baseline IHF%; n=269). (C) Illustrative MRI: a comparison of two male participants, similar age (46 years) and similar baseline WC (105 cm). Participant A was randomly assigned to the MED groups; participant B was assigned to the green-MED group. Both participants lost about 12% of their initial weight after 18 months and reported consuming at least 5–6 time/week walnuts (reported on 28 g/time). Total plasma polyphenol levels at the end of the intervention were higher in the green-MED participant versus MED participant (0.67 mg/L vs 0.24 mg/L). *Significant within-group change versus baseline at 0.05 level. Colour liver images were generated using pride software (by Philips). HDG, healthy dietary guidelines; IHF, intrahepatic fat; ITT, intention to treat; MED, Mediterranean; WC, waist circumference.

Figure 3

Changes in IHF across tertiles/categories of dietary components. Mankai shake and green tea tertiles are calculated from the weighted mean of consumption reported after 6 and 18 months of intervention. serum folate tertiles (of 18-month change in serum folate): T1≤−0.41; T2=−0.40 to 1.46; T3≥1.47; Mankai shake tertiles: T1≤1.67/week; T2=1.68 to 3.00/week; T3≥3.01/week; green tea tertiles: T1≤2/day; T2=2.01 to 3.67/day; T3≥3.68/day; walnut consumption categories: low: 0 to 1–3 times/month; medium: 1–2/week to 3–4/week; high: more than 5–6/week. Categories intervention group distribution for walnuts: low consumption: 60% MED, 40% green-MED; medium consumption: 45% MED, 55% green-MED; high consumption: 45% Med, 55% green-MED. Specific between tertiles/consumption group p values are corrected for multiple comparisons. # none of the participants reported on more processed meat. IHF, intrahepatic fat; MED, Mediterranean; T1, lowest tertile; T2, intermediate tertile; T3, highest tertile.

Figure 4

(A–D) Intrahepatic fat and the gut microbiome. (A) Gut microbiome composition (beta diversity) and IHF% at baseline. Gut microbiome composition and IHF, shown by principal coordinate analysis (PCoA) of UniFrac distances between all baseline samples. Colours denotes 1st (grey) 2nd (yellow) and 3rd (brown) IHF% tertiles. 95% SE ellipses are shown for each tertile. Boxplots on the right describe PCo1 score by IHF% tertile. (B) Gut microbiome composition change and IHF% change. Correlation between principal component 5 (PCo5), the principal coordinate most highly correlated with IHF change (Y axis), and 18-month change in intrahepatic fat. Colours denotes lifestyle intervention group allocation. Boxplots on the right describe PCo5 score by IHF% lifestyle intervention group. (C) Mediation analysis: assessing the proportional mediatory effect of microbiome composition change (measured as PCo5) in the association between lifestyle intervention and IHF% change. (D) Stepwise identification of genus level bacteria associated with: IHF% at baseline (top, two selected bacteria), IHF% 18-month change (middle, heatmap) and with lifestyle intervention (bottom, bar plot, selected bacteria). IHF, intrahepatic fat.