Manguso 2005.

Study characteristics
Methods	RCT. Cross‐over design with 2 treatment arms. Duration: 6 months in total but only results from the first 3 months will be considered. Location: Naples, Italy.
Participants	Inclusion criteria: adults admitted to a specialist unit for the management of liver cirrhosis (Child A or B class). Exclusion criteria: HBV infection, autoimmune liver disease, drug or alcohol use, hepatocellular carcinoma, HIV infection, liver transplantation, impaired renal function, sepsis, or throid dysfunction, following specific diets, ascites and in current or previous treatment with albumin. Number randomised: 90 participants (intervention group, n = 45; control group, n = 45). Attrition: 3 withdrawals, but information from authors, no deaths. Gender split: 52 males and 38 females. Age: mean (IQR) 60 (9) years. Nutritional status: Mean (SD) BMI, intervention 27.8 (2.1) kg/m2; control 28.5 (3.2) kg/m2.
Interventions	Intervention: participants received dietary advice in the form of a controlled diet consisting of specific prescription for macronutrients and calcium. Control: participants received no dietary advice in the form of spontaneous diet.
Outcomes	Survival*, weight*, MAMC*, TSF*, energy intake*, Childs Score, biochemistry profile.
Publication details	Language: English. Funding: none declared. Publication status: peer‐reviewed journal.
Notes	Weight may be inappropriate in analysis due to possible presence of ascites.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated sequence.
Allocation concealment (selection bias)	Low risk	Sequentially‐numbered opaque sealed envelopes.
Blinding (performance bias and detection bias) Clinical outcomes	Low risk	Blinding not described but assessment of mortality unlikely to be influenced by lack of blinding.
Blinding (performance bias and detection bias) Functional outcomes	Unclear risk	None measured.
Blinding (performance bias and detection bias) Nutritional outcomes	Low risk	Information from author, the assessor was blinded to intervention group.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded and likely that researchers and participants were aware of group allocation as this was a nutritional intervention. It is possible that assessment of some outcomes was influenced by lack of blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	The main outcomes assessed were nutritional intake and body composition and information from the author indicated that these were assessed by a researcher blinded to the intervention. Judged unclear as lack of blinding of performance might have influenced nutritional intake.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Information from author: no deaths occurred in the study, 1/45 participant withdrew from the dietary advice group and 2/45 from the no intervention group. 3 additional participants not included because they developed ascites.
Selective reporting (reporting bias)	Unclear risk	No protocol identified. Results on all outcomes specified in the methods reported but as mean (SD) at baseline and end of intervention. Mean change (SD) for weight, energy intake, TSF and MAMC and additional information obtained from authors.
Other bias	Low risk	Baseline characteristics comparable between groups.