Koizumi 2013.

Methods	Japanese multicenter RCT (14 centers in Japan), Phase II
Participants	Number of participants: 93 (December 2008 to February 2012) Number randomized: Experimental group (TSU‐68 + S‐1/CDDP): 46 Control group (S‐1/CDDP): 47 Number evaluated: Experimental group: 45, age (median and range): 62 (30 ‐ 74) years Control group: 46, age (median and range): 63.5 (44 ‐ 76) years Diagnosis: patients with chemotherapy‐naïve unresectable or recurrent AGCs Inclusion: eligible patients should: be > 20 years; with histologically‐ or cytologically‐confirmed adenocarcinoma; unresectable or recurrent gastric cancer; no prior systemic treatment. Recurrent patients were eligible if the last dose of postoperative adjuvant chemotherapy had been received at least 180 days before the start of the study
Interventions	Experimental group : TSU‐68: 800 mg m–2 twice a day day 1 – 35 every 5 weeks; S‐1: 40 – 60 mg m–2 twice a day day 1 – 21 every 5 weeks; CDDP: 60 mg m–2 i.v. on day 8 Control group: S‐1: 40 – 60 mg m–2 twice a day day 1 – 21 every 5 weeks; CDDP: 60 mg m–2 i.v. on day 8 The treatments were continued until 1 of the following occurred: progressive disease (PD), unacceptable toxicity, withdrawal of participant consent (regardless of toxicity), or termination of treatment at the discretion of the attending physician
Outcomes	Duration of follow‐up (median): not mentioned PFS, OS, overall response, adverse events (reported by different symptoms, no data for any adverse event)
Notes	This trial was supported by Taiho Pharmaceutical Co., Ltd
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization was performed according to the minimization method
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding of participants and personnel (performance bias) All outcomes	High risk	The open‐label design may introduce performance bias (e.g. through stress‐related mechanisms)
Blinding of outcome assessment (detection bias) OS, Serious adverse events	Low risk	These outcomes were unlikely to be affected by the blinding status of assessors
Blinding of outcome assessment (detection bias) PFS, Response, quality of life & adverse events	Low risk	All measured images were assessed by a central imaging review committee
Incomplete outcome data (attrition bias) All outcomes	Low risk	91 randomized participants were included in analysis (97.8%)
Selective reporting (reporting bias)	Low risk	No prior protocol was available but all important outcomes were reported
Other bias	Low risk	No obvious potential source of bias