Ishido 2016.

Study characteristics
Methods	Randomised, cross‐over, phase 2 trial with 2 arms comparison of aprepitant + granisetron + dexamethasone vs palonosetron + dexamethasone Recruitment period: November 2010 to August 2013 85 patients enrolled 84 patients evaluated for efficacy analysis (1 patient died) Masking: open‐label Baseline patient characteristics: reported Follow‐up: yes
Participants	Inclusion criteria 20 to 79 years old advanced or recurrent oesophageal or gastric carcinoma chemotherapy‐naïve 2 or more courses of chemotherapy including cisplatin ≥ 60 mg/m² Exclusion criteria no previous chemotherapy serious heart disease, serious renal disease, serious liver disease, poorly controlled diabetes woman during pregnancy or with possibility of pregnancy severe mental disorder allergic past history for serotonin receptor antagonist nausea, vomiting due to brain tumour or ileus planning to receive radiotherapy for chest, abdomen, or pelvis using antiemetic drug within 48 h before chemotherapy judged by the investigator as inappropriate for study entry Median age (range), years: 65 (30 to 75) in aprepitant + granisetron + dexamethasone group, 64 (33 to 77) in palonosetron + dexamethasone group Gender: male + female Tumour/cancer type: advanced or recurrent oesophageal or gastric cancer Chemotherapy regimen S‐1 and cisplatin (SP) S‐1, cisplatin, and docetaxel (DCS) docetaxel, cisplatin, and 5‐fluorouracil (DCF) capecitabine, cisplatin, and trastuzumab (XPT) Country: Japan (single centre)
Interventions	Cross‐over trial Experimental: arm A: aprepitant + granisetron + dexamethasone, then palonosetron + dexamethasone 1 h before start of treatment with cisplatin: 125 mg aprepitant (administered p.o.) + 3 mg granisetron (administered i.v.) + 6.6 mg dexamethasone (administered i.v.) after 24 h and 48 h: 80 mg aprepitant (administered p.o.) + 4 mg dexamethasone (administered p.o.) during second cycle, study treatments were crossed over, that is, aprepitant + granisetron + dexamethasone group received palonosetron + dexamethasone Experimental: arm B: palonosetron + dexamethasone, then aprepitant + granisetron + dexamethasone before treatment with cisplatin: 0.75 mg palonosetron (administered i.v.) + 13.2 mg dexamethasone (administered i.v.) after 24 h and 48 h: 8 mg dexamethasone (administered p.o.) during second cycle, study treatments were crossed over, that is, palonosetron + dexamethasone group received aprepitant + granisetron + dexamethasone
Outcomes	Primary endpoint complete response within 120 h after start of the first course of chemotherapy Secondary endpoints incidences of nausea and vomiting developing within 120 h and proportion of patients who received rescue medication during the first cycle patient preference quality of life assessed on the basis of the FLIE questionnaire adverse events food intake status
Notes	study registered with University Hospital Medical Information Network Clinical Trials Registry (UMIN‐CTR) of Japan (ID UMIN 000005623) self‐funded by Kitasato University School of Medicine, Department of Gastroenterology conflicts of interest: "there are no conflicts of interest"
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients scheduled to receive chemotherapy who provided informed consent were assigned randomly to receive AGD or PD in a 1: 1 ratio"
Allocation concealment (selection bias)	Unclear risk	Comment: allocation concealment not reported
Blinding of participants and personnel (performance bias) Blinding of participants	High risk	Quote: "... open ‐no one is blinded ..."
Blinding of participants and personnel (performance bias) Blinding of personnel	High risk	Quote: "... open ‐no one is blinded ..."
Blinding of outcome assessment (detection bias) Subjective outcomes (Patient reported outcomes)	High risk	Comment: patients and personnel were not blinded towards the intervention and therefore might influence subjective outcomes analysis
Blinding of outcome assessment (detection bias) Objective outcomes (including mortality and safety)	Low risk	Comment: although this was an open‐label study, both patients and personnel had no influence on objective outcomes (e.g. hiccups)
Incomplete outcome data (attrition bias) Subjective outcomes (Patient reported outcomes)	Low risk	Quote: "effectiveness and safety were evaluated in the remaining 84 patients ..."
Incomplete outcome data (attrition bias) Objective outcomes (including mortality and safety data)	Low risk	Quote: "effectiveness and safety were evaluated in the remaining 84 patients ..."
Selective reporting (reporting bias)	Low risk	Comment: all outcome measures were described in the results section
Other bias	Low risk	Comment: no information to suggest other sources of bias