Murphy 1992.

Methods	Randomised controlled trial. Duration 16 weeks. Outcomes not assessed blind.
Participants	Adults (n = 22, all men, mean age (SD) 37.3 years (6.7 years) with HIV infection who had involuntary weight loss of >5%. 6 dropouts, 5 in the dietary counselling group and in the dietary counselling and supplement group.
Interventions	Dietary counselling verbally and in writing to consume a calculated amount of energy and protein per day (n = 11) versus dietary counselling (as above) and 2x 235 ml of a supplement (1.5 kcal/ml) (n = 11).
Outcomes	Survival*, body weight*, BMI*, MUAC*, serum albumin, energy* and protein intake.
Notes	Additional data and information on quality obtained from authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants were selected consecutively as they presented with weight loss. No details of the group for the first participant.
Allocation concealment (selection bias)	High risk	Investigators used alternate allocation.
Blinding (performance bias and detection bias) Clinical outcomes	High risk	Paper states outcomes not assessed blind.
Blinding (performance bias and detection bias) Functional outcomes	High risk	Paper states outcomes not assessed blind.
Blinding (performance bias and detection bias) Nutritional outcomes	High risk	Paper states outcomes not assessed blind.
Incomplete outcome data (attrition bias) All outcomes	Low risk	6 dropouts, 5 in the dietary counselling group and 1 in the dietary counselling and supplement group. 5 dropouts because of subsequent GI disease, and 1 due to self exclusion.
Selective reporting (reporting bias)	Unclear risk	All specified outcomes reported but continuous variables presented as mean (SD) at baseline and end of intervention. Mean change with SDs for weight has been imputed. Data on change in energy intake are presented with precise P values and so mean change (SD) obtained by calculation. Additional data and information on quality obtained from authors.
Other bias	High risk	Baseline variables given, dietary counselling group and weighed 5 kg less than the dietary counselling group and supplement group.