BACKGROUND & AIMS: Non-alcoholic fatty liver disease (NAFLD) affects up to 30% of the population and signifies increased risk of liver fibrosis and cirrhosis, type 2 diabetes, and cardiovascular disease. Therapies are limited. Weight loss is of benefit but is difficult to maintain. We aimed at examining the effect of the Mediterranean diet (MD), a diet high in monounsaturated fatty acids, on steatosis and insulin sensitivity, using gold standard techniques. METHODS: Twelve non-diabetic subjects (6 Females/6 Males) with biopsy-proven NAFLD were recruited for a randomised, cross-over 6-week dietary intervention study. All subjects undertook both the MD and a control diet, a low fat-high carbohydrate diet (LF/HCD), in random order with a 6-week wash-out period in- between. Insulin sensitivity was determined with a 3-h hyperinsulinemic-euglycemic clamp study and hepatic steatosis was assessed with localized magnetic resonance (1)H spectroscopy ((1)H-MRS). RESULTS: At baseline, subjects were abdominally obese with elevated fasting concentrations of glucose, insulin, triglycerides, ALT, and GGT. Insulin sensitivity at baseline was low (M=2.7 ± 1.0 mg/kg/min(-1)). Mean weight loss was not different between the two diets (p=0.22). There was a significant relative reduction in hepatic steatosis after the MD compared with the LF/HCD: 39 ± 4% versus 7 ± 3%, as measured by (1)H-MRS (p=0.012). Insulin sensitivity improved with the MD, whereas after the LF/HCD there was no change (p=0.03 between diets). CONCLUSIONS: Even without weight loss, MD reduces liver steatosis and improves insulin sensitivity in an insulin-resistant population with NAFLD, compared to current dietary advice. This diet should be further investigated in subjects with NAFLD.