Lai 2012a.

Methods	Study design: parallel RCT Study duration: commenced July 2003 Follow‐up period: average 3.6 years
Participants	Country: Hong Kong Setting: multicentre (4 sites) Incident adult patients on CAPD Number: treatment group (58); control group (67) Mean age ± SD (years): treatment group (56.4 ± 1.6); control group (59.5 ± 1.35) Sex (M/F): treatment group (36/22); control group (33/34) Exclusion criteria: malignancy; systemic lupus erythematosus; chronic valvular or congenital heart disease
Interventions	Treatment group low‐GDP PDFs, Gambrosol Trio, Physioneal 40 and Balance Control group Conventional PD Solutions ‐ lactate‐buffered glucose based Dianeal PD‐2 or ANDY‐Disc
Outcomes	Composite co‐primary outcomes biochemical profile of cytokines, growth factors, adipokines, and cardiac biomarkers determined after stable PD treatment for an average duration of 2.3 years dialysis adequacy determined by GFR and daily urine output at initiation and at the time of census after stable PD for an average duration of 3.6 years Also determined UF Urine volume Dialysis Adequacy (Kt/V, CrCl) Residual GFR 4‐hour dialysate:plasma creatinine
Notes	Study supported in part by a Renal Discoveries‐International Society of Nephrology grant and a Baxter extramural grant Randomisation and recruitment into this study unusual. Patients were informed only at 2.3 years after starting their 'study' of their participation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Random assignments were made by the patient's training nursing officer at the individual renal centre
Allocation concealment (selection bias)	High risk	Random assignments were made by the patient's training nursing officer at the individual renal centre
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Open‐label, however unlikely to have affected the objective outcomes measured
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Per protocol analysis
Selective reporting (reporting bias)	High risk	Peritonitis not reported
Other bias	High risk	Although baseline characteristics were reported to be similar. The paper did not disclose the duration of PD that these patients received, so one cannot exclude that they may represent different vintage. Also, for biochemical analyses, there is no baseline value available, thus it is difficult to be certain whether differences are present truly or due to type I error Multiple types of PD solutions used in intervention and control groups.