**3. Results**

The study population consists of 2026 people with T2D (1136 men and 890 women) with a mean age of 62.1 ± 6.5 years, a mean BMI of 30.3 ± 4.5 kg/m<sup>2</sup> and a mean duration of diabetes of 8.5 ± 5.7 years. The prevalence of NASH according to the ION was 32%.

In Table 1 are reported the general characteristics and the cardio-metabolic profile for participants in the two clusters. The BMI, waist and hip circumference, waist/hip ratio, systolic and diastolic blood pressure, HbA1c, fasting plasma glucose and insulin, HOMA-IR, plasma LDL-cholesterol, plasma triglycerides, and liver enzymes were significantly higher among people in the cluster NASH as compared with those in the cluster NO-NASH, while age, diabetes duration and plasma HDL-cholesterol values were significantly lower. The proportion of smokers was similar in the two clusters, and a high proportion of the population was taking lipid lowering medications (62%) and/or antihypertensive medications (93%), with no significant differences between the two clusters.

In Table 2 are reported the energy intake and the nutrient composition of the diet in the participants in the two clusters. A significantly lower intake of energy, fibre, vitamin C, ß-carotene, vitamin E and polyphenols, was observed in the cluster NASH; accordingly the antioxidant capacity of the diet, estimated as Trolox Equivalent Antioxidant Capacity (TEAC), Total Radical-Trapping Antioxidant Parameter (TRAP), Ferric Reducing-Antioxidant Power (FRAP) was significantly lower. No differences were detected for the other components of the diet between the groups.


**Table 1.** General characteristics and metabolic profile in the clusters NASH or NO-NASH.

Data are means ± SD. BMI: body mass index; HbA1c: glycated hemoglobin; HOMA-IR, homeostatic model assessment; eGFR: estimated glomerular filtration rate; AST: aspartate aminotransferase; ALT: alanine aminotransferase; GGT: gamma-glutamyl-transpeptidase.



Data are means ± SD. TEI: total energy intake; TEAC: trolox equivalent antioxidant capacity; TRAP: total radical-trapping antioxidant parameter; FRAP: ferric reducing-antioxidant power.

Coherent with these findings, the rMED score, an indicator of the overall quality of the adherence to the Mediterranean dietary pattern, was significantly lower in the cluster NASH (Table 3). This data was due to significant differences in individual foods and food group consumption (Table 3). People in the cluster NASH showed a significantly higher consumption of pasta, rice, potatoes, total meat, red meat, processed meat and a lower consumption of wholegrain bread, legumes and nuts, vegetables, fruits, white meat, fatty fish, total dairy products, milk and yogurt. The consumption of total cereals, white bread, total fish, lean fish, cheese, eggs, vegetable oils and fats from animal origin was not significantly different between the groups.


**Table 3.** Consumption of specific food items (g/1000 Kcal/day), in the clusters NASH or NO-NASH.

Data are means ± SD. rMED: relative Mediterranean Diet.

#### **4. Discussion**

To the best of our knowledge, this is the first large cross-sectional study evaluating the association between NASH and habitual diet in a well-characterized sample of adults with T2D.

The prevalence of NASH in the study population was high (32%) and in line with recent epidemiological data, as was the more adverse cardiovascular risk factors profile observed in the cluster NASH [12]. The association between nutrient composition of the diet and NAFLD, particularly NAFL, has been reported in prior epidemiological studies [41], but less is known regarding the association with more advanced stages of NAFLD (i.e., NASH) and no data are available on the association between habitual diet and NASH in patients with T2D. Here, we provide data on this association, thus expanding current knowledge.

The first remarkable finding is that the nutrient distribution was largely similar in the NASH or NO-NASH cluster. This is at variance with studies assessing the relation between NAFDL and diet composition in non-diabetic people which describe higher intakes of cholesterol, saturated fat [42,43] and added sugars [44,45] in individuals with NAFLD in comparison with matched controls without NAFLD. This inconsistency might be related to differences in the study design; in fact, we studied a population consisting of patients with T2D regularly attending diabetes clinics and who, therefore, may have restricted the consumption of the above-mentioned items as a result of medical advice, thus diluting the association.

The second relevant finding of this study is the association between fibre, micronutrient intake and NASH. Interestingly, despite the composition in macronutrients being similar between groups, the overall quality of the diet was very different between groups; with this regard, the intake of fibre, vitamins and polyphenols was significantly lower in the cluster NASH. These results are in line with epidemiological and clinical studies performed in patients without T2D [23,46,47]. All together, available evidence suggests that vitamins and polyphenols might prevent the advance of steatosis to NASH, probably restoring oxidative stress and reducing the transcription of the pro-inflammatory cytokines, which are the main drivers in the progression from NAFL to NASH [24,48,49]. Furthermore, dietary fibre might positively influence NAFL by acting on postprandial metabolic state, decreasing glucose absorption with a consequent reduction of the hepatic influx of glucose and de-novo lipo-genesis [22]. In addition, dietary fibre might stimulate a healthy gut microbiota, consequently decreasing the development of tissue inflammation and liver injury that led to NASH [50,51] and expanding also to NASH the interest in bioactive food compounds for T2D.

In terms of food groups, the cluster NASH was characterized by a lower consumption of whole grain bread, legumes and nuts, vegetables, fruits, fatty fish, milk and yogurt and a higher consumption of pasta, rice, potatoes, red meat and processed meat, providing data on the relation between food groups, NASH and T2D in line with those available for people without diabetes [43,52,53].

These data indicate that overall the NO-NASH dietary pattern is close to the Mediterranean dietary model, contributing to the growing evidence suggesting this model as the reference nutritional pattern to prevent and treat NAFLD [14,54], also in people with T2D. The beneficial effects of the Mediterranean diet on NAFLD might be related to dietary components such as dietary fibre, polyphenols and vitamins, that lead to the enhancement in the most important risk factors of NAFLD, such as BMI, insulin resistance and serum triglycerides [54], which are also key pathogenic factors for the development of T2D and major determinants of blood glucose control once diabetes has developed.

To our knowledge, this is the first epidemiological study on a large population of patients with T2D to evaluate the association between NASH and different dietary components and food groups, in a real-life setting.

There are still many open questions with respect to dietary treatment in individuals with both NASH and T2D. Importantly, in spite of the intimate relation between NAFLD and T2D, there are few nutritional intervention trials in which patients with coexisting T2D and NASH have been included. Therefore, it is unclear whether results from the numerous trials performed in patients with NASH and without T2D can be generalized to patients with both diseases. This study, by including a detailed analysis of both vegetable-based and animal-based foods, raises hypotheses on the overall dietary approach for NASH in people with diabetes, which could require confirmation in future large randomized-controlled trials.

Some limitations of the study must be acknowledged. First, the causal relationship between dietary components, Mediterranean dietary score and NASH cannot be proven due to the cross-sectional study design. Second, potential confounding from unmeasured lifestyle factors, such as physical activity level, might exist. Furthermore, data regarding dietary habits were collected only once and, consequently, could be prone to seasonal fluctuation and recall bias. Finally, NASH was detected by an indirect index currently accepted by NAFLD guidelines [16,55]. This index, although not specifically developed for people with diabetes was, however, validated in people with obesity who share several metabolic and clinical features with T2D (obesity, excess of visceral fat, insulin resistance and high prevalence of NASH), [15].

These limitations are counterbalanced by several strengths: a large sample size, a well-defined population of patients with T2D studied within the context of real-life clinical practice, the collection of nutritional and clinical data according to standard methods and biochemical measurements performed in a centralized laboratory.

#### **5. Conclusions**

In conclusion, this is one of the first epidemiological studies to investigate the dietary correlates of NASH in free living people with T2D focusing on foods and food groups. The results provide insights regarding habitual food consumption and dietary components as correlates of NASH in people with coexisting diabetes, showing that the NO-NASH dietary pattern is characterized by a higher intake of whole grain-based foods, legumes, nuts, fruits, vegetables, fatty fish, milk and yogurt, translating into a higher intake of polyphenols, vitamins and fibre and a higher Mediterranean dietary score. These findings expand current knowledge by highlighting the potential for dietary components in the prevention/treatment of NASH in people with T2D.

**Author Contributions:** Conceptualization, M.V., G.D.P. and O.V.; methodology, M.V., G.C., P.C., L.B., S.S. and V.L.; formal analysis, M.V. and G.D.P.; data curation, M.V., G.D.P. and M.M.; writing—original draft preparation, M.V. and G.D.P.; writing—review and editing, M.M. and O.V.; supervision, O.V. and G.R.; funding acquisition, O.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** The study was supported by the Italian Medicines Agency (AIFA) within the Independent Drug Research Program—contract number FARM6T9CET—and by Diabete Ricerca, the non-profit Research Foundation of the Italian Diabetes Society. The funding agency played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the manuscript; or in the decision to submit the manuscript for publication.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the FEDERICO II UNIVERISTY Ethics Committee (protocol code: 123/08; date of approval: 21 January 2008).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The participation of the patients in the study is gratefully acknowledged. We thank all the investigators and the dietitians in the TOSCA.IT centers for their excellent cooperation. We are also indebted to the administrative personnel of the Italian Diabetes Society (SID) for their support.

**Conflicts of Interest:** The authors declare no conflict of interest.
