Liver Fat Is Reduced by an Isoenergetic MUFA Diet in a Controlled Randomized Study in Type 2 Diabetic Patients

A total of 45 participants (37 men and 8 postmenopausal women) with type 2 diabetes were recruited at the diabetes outpatient clinic at the Department of Internal Medicine of the University Medical School Hospital and were enrolled in the study after giving their written informed consent. Patients were overweight or obese, without body weight changes over the previous 6 months, and in satisfactory blood glucose control with diet or metformin plus diet treatment, had normal fasting plasma triglyceride and cholesterol levels, and were not taking any lipid-lowering drugs. Subjects had no evidence of A, B, or C virus or autoimmune hepatitis; clinical signs or symptoms of inborn errors of metabolism; history of toxins or drugs known to induce hepatitis; or history or symptoms of any known disease, apart from diabetes. Their alcohol consumption was <20 g/day. They were not vegetarians or engaged in intensive physical activity. Of the 45 participants enrolled, 4 (1 in each group) dropped out before the final examination as they moved to other cities or could no longer accomplish their working routine. Moreover, in spite of their initial consent, five subjects refused proton nuclear magnetic resonance (¹H NMR) spectroscopy examination because of claustrophobia. Data in this article are from the 36 participants who underwent hepatic fat measurement. The study protocol was approved by the Federico II University Ethics Committee.

In order to detect a 30% difference in liver fat content between treatments with a 80% power at 5% significance level, 40 patients had to be studied. The expected changes in liver fat content after treatment correspond to the differences observed between obese patients with or without type 2 diabetes in a previous study (1).

The intervention was preceded by a run-in period of 3 weeks during which participants were stabilized on a diet with a composition similar to the one usually followed, only providing that saturated fatty acids were at least 13% (carbohydrate 48%, total fat 33%, saturated fat 13%, and protein 18% of total energy intake). With use of a factorial 2 × 2 randomized parallel-group design, the participants were randomly assigned to one of the following intervention groups for an 8-week period: 1) high-carbohydrate/high-fiber/low–glycemic index diet (CHO/fiber group), 2) high-MUFA diet (MUFA group), 3) high-carbohydrate/high-fiber/low–glycemic index diet plus physical activity program (CHO/fiber+Ex group), and 4) high-MUFA diet plus physical activity program (MUFA+Ex group). The allocation to the intervention, stratified for BMI, age, sex, and diabetes therapy (only diet or metformin), was randomly performed by a minimization method using MINIM software (www.users.york.ac.uk). The experimental diets were isoenergetic in order to keep body weight constant and differed in macronutrient composition: carbohydrate 52 vs. 40%, fiber 28 vs. 10 g/1,000 kcal, glycemic index (bread reference) 60 vs. 95%, fat 30 vs. 42%, and MUFA 16 vs. 28% for CHO/fiber diet and MUFA diet, respectively. The other components including the saturated fat (7%), protein (18%) and polyunsaturated fatty acid (PUFA) (4%) were similar in the two diets. The dietary macronutrient composition was drawn by the tables of food composition from the Italian National Research Institute for Food and Nutrition (20). For improvement of dietary compliance, patients were seen weekly by an experienced dietitian, who made telephone calls every 2–3 days to ensure that they followed the assigned diet. Moreover, some foods—olive oil for MUFA diet and stick bran for CHO/fiber diet—were provided to the participants. Adherence to the dietary treatments was evaluated by a 4-day food record (two weekend and two working days) completed by the participants every second week.

The structured supervised exercise program was performed at the Cardiac Rehabilitation Centre of the Department of Clinical Medicine, Cardiovascular and Immunological Sciences. Participants exercised on treadmill or cycle ergometer two times per week for 45 min at an intensity corresponding to 70% of their baseline peak Vo₂; the session also included warm-up and cool-down periods. Only the participants assigned to the exercise program received information about the benefits of physical activity and a motivational counseling. At baseline and at the end of the study, all patients underwent a cardiopulmonary stress test. Every second week, a validated questionnaire (7-Day Physical Activity Recall) (21) was administered to all participants to quantify their weekly moderate/vigorous physical activity apart from the structured training program.

All evaluations were performed before and after the 8-week intervention periods by personnel blinded to the assignment. Body weight, height, and waist circumference were measured in the fasting state according to standardized procedures.

Liver fat content was measured by proton magnetic resonance spectroscopy (¹H MRS) performed on a 1.5T MR scanner (Intera; Philips Medical Systems, Best, the Netherlands) equipped for proton spectroscopy acquisitions in the afternoon 8 h after a light morning breakfast. Sagittal, coronal, and axial slices covering the whole liver were preliminarily taken to position the spectroscopy acquisition voxel. A single voxel of 8 cc (2 × 2 × 2 cm) was placed within the right lobe avoiding major blood vessels, intrahepatic bile ducts, and subcutaneous fat tissue. The proton spectrum was acquired using an eight-channel phased array body coil after shimming over the volume of interest using an unsuppressed water PRESS sequence with the following parameters: repetition time 1,200 ms; echo time, 40 ms; 512 data points over 2,500 KHz spectral width, 12 acquisitions. Total spectra acquisition time was 16 s and thus acquired in breath hold to avoid movement artifacts and spectra broadening. Spectra were analyzed using LCModel software (version 6.2-1; http://s-provench-er.com) that fits in vivo metabolite spectra using model resonances acquired under comparable scanning conditions from multiple compounds in standard phantom solutions (22). Concentration values in arbitrary units of water peak (signal of water) and the sum of lipid peak at 1.3, 0.9, and 1.6 ppm (signal of fat) were considered for fat liver quantification, including both methyl and methylene groups of triglyceride molecule. Signal decay was corrected for the different T2 decay of water and fat using mean T2 relaxation times of 50 and 60 ms for water and fat, respectively. Hepatic fat percentage was calculated using the following formula: 100 · Sf ⁄ (Sf + Sw) (23), where Sf is signal of fat and Sw is signal of water. These values represent a relative quantity of water and fat in the volume of interest. To convert these values into absolute concentrations (weight per volume) expressed as percent fat, equations validated by Longo et al. (24) were applied.

All patients underwent cardiopulmonary exercise stress testing on a bicycle ergometer by a ramp protocol of 20 W/min in the afternoon 8 h after a light morning breakfast. The pedaling was kept constant at 55–65 revolutions per minute. Respiratory gas exchange measurements were obtained breath by breath with use of a computerized metabolic cart (Vmax 29C; Sensormedics, Yorba Linda, CA). The ventilation (VE) versus Vco₂ relationship was measured by plotting VE against Vco₂ obtained every 10 s of exercise (VE/Vco₂ slope): both VE and Vco₂ were measured in liters per minute. The VE/Vco₂ slope was calculated as a linear regression function, excluding the nonlinear part of the relationship after the onset of acidotic drive to ventilation.

Total cholesterol, triglyceride, and glucose concentrations were assayed in plasma by enzymatic methods (Roche Molecular Biochemicals, Mannheim, Germany) on an autoanalyzer (Cobas Mira; ABX Diagnostics, Montpellier, France). LDLs were isolated by preparative sequential ultracentrifugation from the 1,006 density bottom for difference from precipitated HDL (25). Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were determined by colorimetric methods according to the International Federation of Clinical Chemistry recommendations. Plasma insulin concentrations were measured by Enzyme Amplified Sensitivity Immuno Assay (Dia-Source, Nivelles, Belgium). Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following formula: fasting glucose (mg/dL) × fasting insulin (µU/mL)/405.

Data are expressed as means ± SD unless otherwise stated. Differences between the four groups at baseline were tested by one-way ANOVA and post hoc analysis with Bonferroni correction for multiple comparisons. The effects of the four interventions were evaluated by two-way repeated-measures ANOVA. Baseline and after-treatment values were included as levels of within-subjects factor (time) and type of diet and presence/absence of physical activity program as between-subjects factors.

Changes in liver fat content were also evaluated as percent changes [(final values − baseline values)/baseline values × 100] in a general linear model taking the change in liver fat as dependent variable, type of diet and presence/absence of physical activity program as fixed factors, and baseline value of liver fat as covariate. A P value <0.05 was considered significant. The statistical analysis was performed according to standard methods using the SPSS software (SPSS/PC; SPSS, Chicago, IL).

As shown in Table 1, the participants allocated to the four intervention groups were comparable for age, BMI, diabetes treatment, fasting plasma glucose concentrations, and HbA_1c levels.

The two diets followed by the patients were, as expected, significantly different for total fat, MUFA, carbohydrate, fiber, glycemic index, and glycemic load (Table 2). The MUFA diet was richer in vitamin E, while the CHO/fiber diet was richer in polyphenols and had a higher oxygen radical absorbance capacity (ORAC) than the MUFA diet (Table 2). Both diets were well accepted by the patients. Participants allocated to CHO/fiber diet were considered compliant with the treatment if carbohydrate intake was at least 50% of total energy, glycemic index <64%, and daily fiber intake at least 40 g; participants allocated to MUFA diet were considered compliant if carbohydrate intake was <45% of total energy, glycemic index was >86%, and MUFA proportion was at least 25% of total energy. All subjects allocated to MUFA diet satisfied the adherence criteria. One patient allocated to CHO/fiber group was considered not compliant, since the glycemic index was higher than accepted (70%). This participant was excluded from the analysis of liver fat change. Body weight and waist circumference did not change significantly during intervention in the four groups (Table 3).

Patients allocated to the exercise groups were considered compliant with the treatment if they attended at least 70% of prescribed training sessions. Because of employment schedule changes, two patients in the MUFA+Ex group and one in the CHO/fiber+Ex group did not satisfy compliance criteria. Results were similar by including or not including these three subjects, and therefore we report results on all participants. There were no differences between groups in moderate/vigorous weekly physical activity not including the supervised exercise program as measured by the 7-Day Physical Activity Recall questionnaire (CHO/fiber group 5.6 ± 3.8 kcal/kg/week, MUFA group 4.9 ± 3.9 kcal/kg/week, CHO/fiber+Ex group 6.5 ± 4.9 kcal/kg/week, and MUFA+Ex group 9.2 ± 10.9 kcal/kg/week; P = 0.541 by one-way ANOVA).

The training program induced a significant increase in Vo₂ peak in the subjects allocated to MUFA+Ex group and not in those in the CHO/fiber+Ex group (Table 3). Results were not changed by excluding noncompliant participants (MUFA+Ex group 16.4 ± 1.8 vs.18.2 ± 2.2 mL/kg/min, P = 0.007, and CHO/fiber+Ex 13.7 ± 3.7 vs.13.9 ± 4.3 mL/kg/min, P = 0.762). So as to maintain stable weights during the intervention, the exercise groups were not required to increase their dietary caloric intake significantly more than the nonexercising groups (181 ± 287 vs. 372 ± 379 kcal/day; P = 0.080).

Fasting plasma glucose did not change at the end of the interventions in all groups. Fasting plasma concentrations of total, LDL, and HDL cholesterol and triglyceride did not change at the end of the interventions. At baseline, AST and ALT levels were not different among the four groups. AST levels decreased significantly after MUFA+Ex treatment, but this variation was not statistically significant from the changes in the other groups. Blood glucose control as shown by HbA_1c levels significantly improved in the MUFA group while did not change in the other groups (P < 0.05 for diet effect by repeated-measures ANOVA). After intervention, HOMA-IR was not significantly different from baseline in all groups (Table 3).

Liver fat content measured by ¹H NMR significantly decreased in the MUFA group (7.4 ± 2.8 vs. 5.2 ± 2.7%; P = 0.01) and MUFA+Ex group (11.6 ± 8.0 vs. 9.1 ± 7.4%; P = 0.02); it did not change significantly in the CHO/fiber group (17.7 ± 9.7 vs. 16 ± 6.8%; P = 0.295) or CHO/fiber+Ex group (8.8 ± 4.9 vs. 8.9 ± 5.7%; P = 0.794) (Table 3). Two-way repeated-measures ANOVA, including baseline values as covariate, showed a significant effect on liver fat content for diet (P = 0.006), with no effects for exercise training (P = 0.789) and diet-exercise interaction (P = 0.712) (Table 3). Measured as percent variations (end values − baseline values × 100/baseline values), hepatic fat was reduced by the MUFA diet with (−25%) or without (−29%) exercise significantly more than by CHO diet with (−6%) or without (−4%) exercise (P < 0.05 by ANOVA) (Fig. 1).

The effects of diet on liver fat content were still more evident combining the groups with the same diet (CHO/fiber and CHO/fiber+Ex vs. MUFA and MUFA+Ex). In the absence of baseline differences in hepatic fat content between the two groups, liver fat did not change significantly after CHO/fiber diet (13.1 ± 8.6 vs.12.3 ± 7.1%; P = 0.382), while it significantly decreased after MUFA diet (9.6 ± 6.3 vs. 7.2 ± 5.9%; P < 0.0001).

This is the first randomized intervention study evaluating the effects of qualitative dietary changes and their interaction with aerobic exercise on liver fat content in type 2 diabetic patients in the absence of weight loss. We evaluated the effects on liver fat of the two most recommended dietary approaches to diabetes, i.e., a relatively high-CHO/high-fiber/low–glycemic index diet and a high-MUFA diet with and without the association of a structured physical activity program. Hepatic fat content significantly decreased after the 8-week intervention with MUFA diet independently of exercise. This decrease was clinically relevant (−29%), indicating that qualitative isocaloric dietary modifications may positively influence fatty liver.

In the current study, the MUFA diet differed from the CHO/fiber diet for the high-fat content (40% of total energy) because of the high contribution of MUFA (28 vs. 16%). In the only published isocaloric dietary intervention on fatty liver in overweight nondiabetic women, a high-fat diet (56%) significantly increased the hepatic fat content compared with a low-fat/high-carbohydrate diet (12). The extreme increase in fat in that study was mainly due to saturated fatty acids. Taken together, the data suggest that both the quantity and the quality of dietary fat have a role in the treatment and possibly in the pathogenesis of fatty liver. Differential effects of the type of fatty acids on direct measures of liver fat content have been reported only in animal studies. In mice without fatty liver, an olive oil–rich diet increased hepatic triacylglycerol content compared with corn oil (26), while in a rat model of NAFLD, olive oil decreased the accumulation of triglyceride in the liver significantly more than fish oil and butter (27).

MUFAs may exert their beneficial effects on hepatic fat content through their influence on lipid metabolism in the liver or the abdominal adipose tissue. Since in the postprandial phase MUFAs are more rapidly oxidized than saturated fatty acids (28), a high-MUFA diet may avoid fat deposition in liver by the activation of catabolic pathways. In addition, MUFAs could favor fatty acid deposition in adipose tissue rather than in liver, enhancing the clearance of circulating triglyceride-rich lipoproteins by lipoprotein lipase. We demonstrated in a previous study (29) that a MUFA-rich diet stimulated the activity of lipoprotein lipase more than a saturated fat–rich diet. There is also evidence that MUFAs are preferentially taken up by adipose tissue and muscles because of their greater affinity for the fatty acid transporters (30). MUFAs may also improve hepatic steatosis by positively influencing adipose tissue cross-talk with liver metabolism through regulation of adipokines and inflammatory markers synthesis. In fact, in obese women hepatic fat content is associated with increased concentrations of ceramide and inflammatory mediators in subcutaneous adipose tissue (31). Moreover, in cultured cells, saturated fatty acids are preferentially stored as diacyglycerol and ceramides in contrast with MUFAs, which are, instead, preferentially incorporated into triacylglycerols (30). The effects of the MUFA diet did not seem to be mediated and/or lead to changes in insulin resistance, since no significant changes in homeostasis model assessment index were observed in our patients. However, this index is not a good marker of overall insulin resistance in patients with diabetes, and in addition, we had no measures of hepatic insulin sensitivity.

In this study, the MUFA diet had a higher content of vitamin E, which has been shown to induce a reduction in hepatic steatosis in obese nondiabetic patients (32), although it did not significantly modify histological features of NAFLD in obese children and adolescents (33). However, it is unlikely that in our study the antioxidant effects of vitamin E were responsible for the effects of the MUFA diet because of the much higher antioxidant capacity in the CHO/fiber diet that included higher ORAC levels and higher concentrations of polyphenols and vitamin C.

In our study, the high-fiber low–glycemic index diet did not increase liver fat content, while an increase in liver fat has been repeatedly associated with high-carbohydrate diets in previous studies using carbohydrates mainly consisting of fructose, simple sugars, or high–glycemic index foods (16). To date, only one study in mice has compared the effects of high– and low–glycemic index diets, showing that hepatic steatosis increased more with rapidly than slowly absorbed carbohydrates (34). This is in line with the fact that, although dietary carbohydrates may stimulate hepatic de novo lipogenesis leading to fasting hypertriglyceridemia and hepatic fat accumulation (35), the eucaloric substitution of dietary starch for sugar reduces hepatic lipogenesis (36). Therefore, carbohydrates with high and low glycemic index may differently influence hepatic fat deposition through a dissimilar influence on fatty acid synthetic pathways in the liver. The relationship between quality of ingested carbohydrates and fatty liver may also be mediated by changes in gut microbiota because of different modalities of intestinal absorption. Intestinal flora is related to obesity and metabolic diseases, and the microbial species hosted in our intestine may influence metabolite absorption and liver fat metabolism (37).

The effects on fatty liver observed in the current study were independent of the exercise training program. In fact, neither physical activity alone nor its interaction with diet influenced hepatic fat content. In the only intervention trial that explored the effects of aerobic physical activity, independent of weight loss, a 20% reduction in liver fat content was observed (18). Differing from our study, that trial was conducted in nondiabetic patients with a higher baseline exercise capacity (Vo₂ peak ≈25 vs. 15 mL/kg/min in our study). A recent study evaluating the effects of exercise training on fructose-induced liver fat accumulation in rats showed that physical activity did not reduce hepatic fat content, although it beneficially modified the partitioning of lipids, providing more unsaturated fatty acids in liver (38). Because of these inconsistencies, more data are needed to elucidate the role of physical activity on liver fat content in type 2 diabetes. It should also be considered that, because of the cultural resistance and structural barriers to physical exercise common in this middle-age population of persons with type 2 diabetes as well as the need for supervised sessions that could be practically inserted in a clinical setting (work schedule, etc.), the physical exercise implemented in this study was of light/moderate intensity with a relatively low volume of exercise. Therefore, we cannot exclude that a more prolonged/vigorous exercise would have induced beneficial effects, also in light of the resistance to the beneficial effects of physical activity reported in type 2 diabetes (39), which could mainly concern the obese patients (40).

This study has some strengths and limitations. This was a randomized controlled trial with a rigorous follow-up of dietary adherence and direct supervision of physical exercise. Hepatic fat content was measured by the gold standard NMR spectroscopy. A limitation is that only patients in good metabolic control were studied, and therefore our results may not be extended to type 2 diabetic patients of other ethnic groups or with poorer control. Nevertheless, there are no indications that worse metabolic control could negatively influence the effects of MUFAs on hepatic fat content. In spite of randomization, one group (CHO/fiber) had a higher fat liver content at baseline, which was not associated with differences in any other variable known to influence liver fat content. Anyhow, this did not influence the results because analyses were corrected for baseline levels, and the significant decrease in liver fat was not observed in the group with higher baseline values.

In conclusion, this study demonstrates for the first time that qualitative dietary changes without weight loss can beneficially influence hepatic fat content in type 2 diabetic patients. These results are clinically relevant and have implications for the nutritional management of fatty liver in type 2 diabetes, suggesting that a MUFA-rich diet also enriched in fiber and with low–glycemic index should be preferred. Further studies are needed to elucidate the possible mechanisms of these dietary effects and to evaluate the effects of more intense and prolonged physical exercise on liver fat in type 2 diabetes.

The work presented here was supported by ETHERPATHS project (European Community contract no. FP7-KBBE-222639). L.B. received a research grant from the Italian Diabetes Society: “Borsa di studio annuale SID-AMD Pasquale Di Coste.”

No potential conflicts of interest relevant to this article were reported.

L.B. wrote the manuscript, designed the experiment, and collected and analyzed data. A.P. performed imaging studies. G.A. wrote the manuscript and designed the experiment. L.C. collected and analyzed data and provided dietary counseling. A.M. performed biochemical measurements. A.V. collected and analyzed data and supervised exercise training. R.M. collected and analyzed data. M.L. and M.M. performed imaging studies. C.V. designed the experiment and supervised exercise training. G.R. designed the experiment. A.A.R. wrote the manuscript and designed the experiment. A.A.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented at the 72nd Scientific Sessions of the American Diabetes Association, Philadelphia, Pennsylvania, 8–12 June 2012.

The authors are very grateful to the staff of the Department of Clinical and Experimental Medicine, Federico II University, in particular to Paola Cipriano for technical laboratory assistance; Angela Giacco, RD, Anna Rivieccio, RD, and Marisa Giorgini, RD, for dietary counseling; and Salvatore Turco, MD, and Ciro Iovine, MD, for patient recruitment.