Rao 2010.

Methods	Multicenter RCT (22 centres in the UK and Europe), Phase II
Participants	Number of participants: 72 (Auguest 2005 to November 2006) Number randomised: Experimental group: 36 Control group: 36 Number evaluated: Experimental group: 35, age (median and range): 59 (29 ‐ 79) years Control group: 36, age (median and range): 64 (36 ‐ 76) years Diagnosis: patients with previously untreated, histopathologically‐confirmed metastatic gastric adenocarcinoma or adenocarcinoma of the lower third of the esophagus; Only patients with EGFR‐positive tumors were enrolled into the study Inclusion: eligible patients should: be > 18 years; ECOG performance status 0/1; normal cardiac function (left ventricular ejection fraction within the institutional normal range); a minimum 12‐month interval from completion of any neoadjuvant or adjuvant chemotherapy; a minimum 4‐week interval from completion of radiotherapy; and adequate liver, bone marrow and renal function
Interventions	Experimental group : 800 mg matuzumab weekly; epirubicin 50 mg/m², cisplatin 60 mg/m² on day 1 and capecitabine 1250 mg/m² daily in a 21‐day cycle Control group: epirubicin 50 mg/m², cisplatin 60 mg/m² on day 1 and capecitabine 1250 mg/m² daily in a 21‐day cycle Treatment cycles were repeated every 3 weeks for a maximum of 8 cycles of ECX unless there was evidence of disease progression or unacceptable toxicity, death occurred or consent was withdrawn. Matuzumab was continued as a single agent after the 8 cycles of ECX unless there was evidence of disease progression or unacceptable toxicity, death occurred or consent was withdrawn
Outcomes	Duration of follow‐up (median): 28.5 months for experimental group, and 23 months for control group OS, PFS, overall response, quality of life, adverse events (reported by different symptoms, no data for any adverse event)
Notes	Sponsor: Merck KGaA The sponsor involved in the study design, data collection, and data analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random sequence was generated by a computer programme
Allocation concealment (selection bias)	Low risk	Randomization was carried out centrally in a blinded manner by telephone using an interactive voice response system
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	Blinded radiological review was performed
Incomplete outcome data (attrition bias) All outcomes	Low risk	71 participants included in the analysis (98.6%); ITT was used
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