Ito 2004.

Study characteristics
Methods	Study design: parallel RCT Study location: Japan Study dates: enrolment February 1991 to March 1993 Recruitment: patients recruited from 93 institutions in the Chubu region of Japan Number screened / eligible: not reported Inclusion criteria: quote: "primary colon cancer with curative resection and positive macroscopic lymph node metastases (macroscopic Dukes’ C); under 75 years of age with PS 0 or 1; no serious complications; no preoperative treatment (radiation therapy, chemotherapy, immunotherapy) for cancer; no previous history of cancer; acceptable preoperative laboratory data (white blood cell count ≥3,000/mm3, platelet count ≥100,000/mm3, glutamic‐oxaloacetic transaminase and glutamic‐pyruvic transaminase ≤100 IU); preoperative records of serum carcinoembryonic antigen (CEA) level (below 5 ng/dl or not) and purified protein derivative (PPD) skin test reactions (smaller than 10 mm or not)” Exclusion criteria:no additional criteria specified Diagnostic criteria for inclusion/exclusion: quote: “primary colon cancer with curative resection and positive macroscopic lymph node metastases (macroscopic Dukes’ C)"
Participants	Number randomised: 441; Coriolus versicolor (PSK) 220; control 221 Baseline imbalances: groups well‐matched on most variables. Quote: "The clinical characteristics of the two study groups were similar at base line (Table 1). However, there was still an imbalance in the distribution of the patients’ performance status (PS) between the two groups. Thus an adjusted analysis with two balancing factors (lymph node metastases, preoperative CEA level) and PS was adopted in the blinded review as a primary analysis." Age:Coriolus versicolor (PSK) mean 60.4 years (87% age 50 or more); Control mean 60.0 (84% age 50 years or more). Overall mean = 60.0 years Sex:Coriolus versicolor (PSK) male n =123 (55.9%); Control male n=121 (54.8%); Overall male n=244 (55%) Cancer type: colon Cancer removed or present: removed If cancer was present,was it local or metastatic: N/A. However note that there were positive lymph node metastases. What was the aim of chemo/radiotherapy (if applicable)? adjuvant. Withdrawals and exclusions:Coriolus versicolor (PSK) group withdrawals from treatment n = 25 (no hospital visits n = 11, adverse events n = 5, concomitant diseases n = 4, other n = 5); control group withdrawals from treatment n = n = 27 (no hospital visits n = 16, adverse events n = 4, concomitant diseases n = 2, others n = 5).
Interventions	Coriolus versicolor intervention: oral protein‐bound polysaccharide K (PSK) . Dose of 3 g/day from 29 days post‐op for 4 weeks, followed by start of oral 5‐FU for 4 weeks. Ten pairs of PSK followed by 5‐FU treatments = 80 weeks of alternating treatment beginning at four weeks post‐op. Control intervention: chemotherapy alone Chemotherapy regimen: 5‐FU IV 1,000 mg/m2/24hr on postoperative days 1‐2, 8, 9, 15, 16, 22 and 23, followed by 4 weeks PSK 3 g/day alternating with 4 weeks 5‐FU oral 200 mg/day for a total of 10 cycles. Planned length of treatment: 80 weeks. Setting: not described but presumably hospital outpatient. Co‐interventions: none mentioned.
Outcomes	1. Survival 2. Disease‐free survival 3. Treatment modification/withdrawal 4. Adverse events   Planned length of follow‐up from treatment initiation: 7 years Time points: Survival and disease‐free survival reported at 5 and 7 year follow‐up. Other outcomes: cancer‐related deaths, survival from cancer‐related death, and laboratory tests, including white blood cell, neutrophil, etc., oxidative stress index using the ROMs Test to measure hydroperoxide concentration, and salivary chromogranin A. Adverse event assessment: no description of how adverse events were assessed. All adverse events are reported according to numbers of each event and to numbers of patients/year experiencing each event, not to numbers of people experiencing each adverse event as of Year 7 or any other time point.
Notes	Study name: Study Group of Immunochemotherapy with PSK for Colon Cancer Funding source: no description of funding. Declaration of interest: none declared.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were allocated to either the PSK group (group P, n=220) or control group (group C, n=221) according to a dynamic balancing method (modified minimization method) using three balancing factors and institution"
Allocation concealment (selection bias)	Low risk	Quote: "based on central‐registration method between 25 and 28 postoperative days. The randomized results processed in a computed system with the intention of concealment and adequate randomization"
Blinding of participants and personnel (performance bias) Survival	Low risk	No blinding however no likely influence of unblinding upon survival.
Blinding of participants and personnel (performance bias) Disease free survival/recurrence/response	Low risk	No blinding however no likely effect upon disease progression or recurrence.
Blinding of participants and personnel (performance bias) Adverse events ‐ subjective	High risk	No blinding and may have affected experience of adverse events.
Blinding of participants and personnel (performance bias) Treatment modification/withdrawal	Unclear risk	No blinding and may have affected treatment decisions.
Blinding of participants and personnel (performance bias) Adverse events ‐ objective	Unclear risk	No blinding and effect upon adverse events unclear.
Blinding of outcome assessment (detection bias) Survival	Low risk	No blinding of outcome assessors reported however low risk of bias in ascertainment of death.
Blinding of outcome assessment (detection bias) Disease free survival/recurrence/response	Unclear risk	No blinding described and unclear how assessment carried out.
Blinding of outcome assessment (detection bias) Adverse events ‐ subjective	High risk	No blinding described and high risk that assessment may have been affected.
Blinding of outcome assessment (detection bias) Treatment modification/withdrawal	Unclear risk	No blinding described and unclear how assessment carried out.
Blinding of outcome assessment (detection bias) Adverse events ‐ objective	Unclear risk	No blinding described and unclear how assessment carried out.
Incomplete outcome data (attrition bias) Survival	Low risk	Appears complete outcome assessment and reporting.
Incomplete outcome data (attrition bias) Disease free survival/recurrence/response	Low risk	Appears complete outcome assessment and reporting.
Incomplete outcome data (attrition bias) Treatment modification/withdrawal	Unclear risk	Appears complete outcome assessment and reporting.
Incomplete outcome data (attrition bias) Adverse events ‐ objective	Unclear risk	Unclear whether complete assessment was conducted and reported.
Incomplete outcome data (attrition bias) Adverse events ‐ subjective	Unclear risk	Unclear whether complete assessment was conducted and reported.
Selective reporting (reporting bias)	Unclear risk	Protocol is mentioned but a registration or publication was not accessible.
Other bias	Low risk	No potential biases were apparent that do not correspond to the above categories (including study‐specific problems such as carryover in cross‐over trials, more general problems in study conduct such as contamination between intervention arms, or that a study is claimed to be fraudulent).