Misciagna 2017.

Methods	RCT of parallel‐group design
Participants	Participants with non‐alcoholic fatty liver disease (NAFLD) were identified during the NutriEp survey enrolment process (Puglia, Italy). Eligible participants were those with moderate or severe NAFLD (N = 203). Exclusion criteria included: overt cardiovascular disease and revascularisation procedures; stroke; clinical peripheral artery disease; T2DM; more than 20 g/daily of alcohol intake; severe medical condition that may impair the person participating in a nutritional intervention study; people following a special diet or involved in a programme for weight loss, or who had experienced recent weight loss and inability to follow a Mediterranean diet for religious or other reasons 98 participants randomised; 50% men
Interventions	Intervention Low glycaemic index Mediterranean diet (LGIMD). Foods in LGIMD have all a low glycaemic index (GI) and no more than 10% of total daily calories coming from saturated fats. The LGIMD was high in monounsaturated fatty acids from olive oil and contained also omega 3 polyunsaturated fatty acids, from both plant and marine sources. Adherence to the LGIMD as measured by Mediterranean Adequacy Index. Control Italian National Research Institute for Foods and Nutrition (INRAN) guidelines. The recommended diets were provided in brochure format, with graphical explanations organised according to a traffic light system: with a list of foods that can be consumed frequently (green foods), sometimes (yellow foods) and never (red foods). The brochure also contained a dietary record, where participants daily indicated the code of each food consumed at breakfast, lunch, dinner and during snack time. Monthly follow‐up visits in both groups included a face‐to‐face interview with the dietician in order to assess the diet followed by the subject and to give, if needed, personal recommendations to achieve the "group assigned" goal. 6 months follow‐up
Outcomes	Total cholesterol, HDL cholesterol, triglycerides
Notes	Data provided as number and percentage of participants with normal and altered levels rather than mean and SD at baseline and follow‐up so these cannot be used in meta‐analyses. Findings are reported narratively in text.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned, according to a computerised random number sequence
Allocation concealment (selection bias)	Unclear risk	With the exception of the dietitians, investigators and staff were unaware of the participants' diet assignment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	States dieticians were aware of group assignment. States blinding and equipoise were strictly maintained by emphasising to the intervention staff and participants that each diet adhered to healthy principles. Blinding of participants and personnel for behavioural interventions is difficult and often not possible so we have not judged this as at high risk of bias.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Staff members who obtained outcome measurements were not informed about diet assignment
Incomplete outcome data (attrition bias) All outcomes	Low risk	The primary analysis was intention‐to‐treat. 6/50 individuals were lost in the follow‐up in the intervention group and 2/48 in a control group.
Selective reporting (reporting bias)	Low risk	All outcomes stated are reported
Other bias	Unclear risk	Insufficient information to judge