Impact of a Mediterranean Diet on Hepatic and Metabolic Outcomes in Non-alcoholic Fatty Liver Disease

The Medina Randomised Controlled Trial

Elena S. George; Anjana Reddy; Amanda J. Nicoll; Marno C. Ryan; Catherine Itsiopoulos; Gavin Abbott; Nathan A. Johnson; Siddharth Sood; Stuart K. Roberts; Audrey C. Tierney


Liver International. 2022;42(6):1308-1322. 

In This Article


The aim of this RCT was to determine whether the MedDiet was more beneficial compared to an LFD in patients with NAFLD in relation to several metabolic parameters including intrahepatic lipids and insulin resistance, as well as LSM and other parameters of health. Furthermore, the study aimed to assess the relationship between changes in the above metabolic factors and body composition across the two dietary patterns. Importantly, the results of this RCT showed that MedDiet did not significantly reduce hepatic steatosis and was not superior to LFD in an Australian free-living NAFLD cohort. Indeed, we found the LFD significantly reduced hepatic steatosis and HOMA-IR in NAFLD subjects following the 12-week intervention, in contrast to MedDiet that led to non-significant albeit clinically relevant, reductions in both metabolic parameters. These changes however were not significant when compared between-groups. However, an important finding from the study was that both dietary interventions resulted in significant reductions in visceral fat, with the reduction being more pronounced in the LFD group.

These findings align with evidence that weight loss elicits improvements in hepatic steatosis with the LFD group achieving almost 4% weight loss, in 12 weeks despite the intervention having an emphasis on weight maintenance. In contrast, although there were potentially, clinically meaningful improvements in hepatic steatosis in the MedDiet group despite remaining essentially weight stable, these were not statistically significant. This reiterates that energy restriction and subsequently weight loss appear to play a central role in the management of NAFLD. However, weight loss is difficult to achieve and hard to maintain and what this study indicates is that changes in dietary behaviours such as choosing foods that improve dietary quality may be sustainable and lead to small but meaningful effects in NAFLD. Still, current EASL–EASD–EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease recommend the MedDiet as the optimal dietary pattern.[35] In comparison, the American Association for the Study of Liver Diseases (AASLD) guidelines are more conservative recommending that additional rigorous, prospective studies such as ours and longer-term data with histological endpoints are required before specific macronutrient compositions can be recommended.[36] Although the AASLD wording only acknowledges macronutrients, dietary patterns which represent the whole diet, are needed to make sustainable lifestyle changes. Other guidelines for NAFLD agree with those of AASLD and EASL, APASL clearly highlights the importance of a hypocaloric diet. It is important to note that such guidelines are based on studies which are predominantly conducted in Mediterranean countries, and it seems the application of MedDiet in a multiethnic, diverse Western population such as in the present study is less straightforward. For example, to better understand the context and implications of adding energy-dense foods to a multiethnic diet and how this can be balanced ad libitum or to achieve a caloric deficit. Furthermore, issues such as length of dietary intervention and additional participant support(s) to achieve higher levels of compliance are needed in multiethnic non-Mediterranean populations to obtain the full translational benefits of a MedDiet.

There have been few studies showing significant beneficial effects with the provision of MedDiet in Western populations with NAFLD. Ryan et al conducted a cross over RCT under tightly controlled conditions including the full provision of diet.[14] This important proof of concept study showed that the MedDiet does indeed improve hepatic and metabolic outcomes in participants with NAFLD. However, under less stringent conditions, as shown in the present RCT in free-living adults, it appears that more attention is warranted on how to apply and therefore translate the MedDiet into a population who does not habitually follow this dietary pattern. Furthermore, the addition of extra virgin olive oil high in mono-unsaturated fatty acids, antioxidants, fat-soluble vitamins and polyphenols is believed to be one of the key elements providing health benefits in the MedDiet, and yet is energy dense, and therefore adding it to a predominantly Western Diet can lead to higher overall energy intake, as seen in this study, albeit weight remained stable. Furthermore, the results from this study are reinforced by another study conducted in Western Australia where ad libitum MedDiet and LFD both improved hepatic steatosis indicating improvement in overall diet quality regardless of dietary prescription was beneficial.[15] The present study where the LFD group achieved almost 4% weight reduction contrasts with the former study where no changes in weight were observed. While the reduction in total energy consumption did not reduce significantly within the LFD the change is noteworthy and with adequate power for secondary outcomes such a reduction may have reached significance. Furthermore, a modest reduction in total fat intake in the LFD group shows a further improvement in diet quality at the macronutrient level.

In the LFD group, there were meaningful reductions in weight, waist circumference and BMI. However, this study was not sufficiently powered for these secondary outcomes, and thus these changes were not statistically significant. Nonetheless robust evidence from a meta-analysis of eight randomised controlled trials reported that 5% weight loss has been shown to improve hepatic outcomes.[37] This meta-analysis is in agreement with the results from this study for the low-fat diet group where weight loss resulted in a significant reduction in hepatic steatosis. The reduction in visceral fat observed in the LFD group was also not surprising given the weight loss achieved. Of interest, however, the MedDiet group also saw a substantial and significant reduction in visceral fat after 12 weeks despite this group remaining largely weight stable. This supports the evidence that MedDiet can elicit favourable changes in body composition in the absence of weight loss.[38] To explain this, another study comparing three ad libitum diets in overweight individuals for 6 months concluded that diet composition had no significant effect on preventing weight regain. The authors did, however, find that a diet rich in fat, especially MUFAs resulted in less body fat accumulation compared to the control diet.[39] This is in line with other studies showing components of a MedDiet including MUFA are inversely associated with abdominal adipose tissue accumulation regardless of body weight.[40] The mechanism for such benefits is largely unknown, however may be, at least in part, explained by the overall improvement in diet quality associated with increased MUFA intake, as many high MUFA foods and nutrient and antioxidant rich. In this small sample, the reduction in visceral fat after 12 weeks of MedDiet did not lead to statistically significant reductions in hepatic steatosis and HOMA-IR although our observation of a clinically relevant 0.5 unit reduction in HOMA-IR is not dissimilar to other studies in which a reduction of 0.8 units in HOMA-IR with metformin and diet was demonstrated over 9 months.[41] While our study was powered to see a decrease in hepatic steatosis with the dietary patterns, the effect size in metabolic parameters between-groups was not large enough to see a significant between-group difference.

Mediterranean diet studies in NAFLD are limited in non-Mediterranean countries such as Australia and our results showing a lack of superiority of MedDiet over LFD need to be carefully considered alongside other studies. For example, the lack of difference could be because the LFD was more familiar as participants may have potentially received advice on this diet before and in addition may have perceived these foods to be more freely available. The literature showing Mediterranean diet is beneficial, while demonstrated several times has not been shown to be superior in free living, Western or indeed Australian cohorts.[15] The changes in dietary adherence, as indicated by the MEDAS score were also marginal, and this too could explain the modest changes seen. However, it does appear that the MedDiet contributed to improving diet quality and in metabolic outcomes such as visceral fat reduction even in the absence of weight loss. Given the paucity of studies assessing the effects of a MedDiet alongside weight loss targets are lacking in NAFLD subjects, prospective studies are needed to determine if there are combined benefits with regard to improved diet quality and weight loss, beyond the LFD. Such studies are needed a priori to inform dietary recommendations and guidelines for NAFLD in the absence of effective pharmaco-therapy.

The main strengths of this study are its randomised controlled trial design which was carried out in free-living adults with NAFLD. Furthermore, the participants all had a matched number of face-to-face and phone appointments, controlling for any biases around contact with a healthcare professional. We also included robust collection of a 3-day food diary to measure dietary compliance which was validated by an accredited practising dietitian who also provided telehealth support between face-to-face appointments. The MEDAS score was also used to measure compliance with MedDiet and LFD. Hepatic outcomes were also robust including the gold standard MRS approach for IHL quantification, as well as Fibroscan for liver stiffness measure. Still, the limitations of this study include its small sample size which despite being powered on hepatic steatosis, with the unintended weight loss in the LFD group the effect size was not sufficient to assess between-group differences. The small sample size also limits generalizability of the secondary outcomes presented herein. Using BIA for assessing visceral fat has shown conflicting evidence with regard to validity and is less accurate in obese individuals and thus higher quality imaging such as CT or MRI should be used to verify these findings in future studies.[42–44] Furthermore, in dietary interventions it is not possible to blind clinicians and participants from dietary prescription. Despite being comparable in duration to similar studies[15,17,45] longer duration (or term) studies are needed to show efficacy, particularly in relation to outcomes such as LSM and to determine the sustainability of dietary interventions and impact on dietary behaviours. The shorter duration may also explain why some parameters which showed clinically meaningful improvement were not statistically significant. Baseline diets in this study cohort were also not in line with the "poor" quality intake often seen in individuals with NAFLD,[46] and future studies may consider screening participants and recruiting those with poor diet quality at baseline. Finally, the baseline differences between study groups, despite randomisation may have had some impact on scope for changes with high glucose, visceral fat and heart rate in the LFD group. Diets overall were better than that of the general Australian population, also potentially hindering the scope to improve diet quality and thus outcomes in only 12 weeks.