RATIONALE: Drugs used in chemotherapy, such as fluorouracil, work in different ways to stop the growth of tumor cells, either by killing the cells or by stopping them from dividing. It is not yet known whether giving fluorouracil into the liver is more effective than no further treatment for patients with colon cancer undergoing surgery.

PURPOSE: This randomized phase III trial is studying giving infusions of fluorouracil into the liver in treating patients with Dukes' A, Dukes' B, or Dukes' C colon cancer undergoing surgery.

OBJECTIVES: Determine whether adjuvant therapy with portal hepatic perfusion of 5-fluorouracil and sodium heparin effectively prolongs the disease-free interval and increases survival in patients undergoing curative resection of adenocarcinoma of the colon.

OUTLINE: Randomized study. Patients are randomized preoperatively; those randomized to Arm 1 begin therapy intraoperatively or within 6 hours of colonic resection. Arm 1: 5-Fluorouracil plus sodium heparin. Those randomized to Arm 2 receive no adjuvant therapy. Arm 2: observation (no further treatment).

PROJECTED ACCRUAL: 1,334 patients will be entered over a 4-year period.

Inclusion Criteria:

• more than one synchronous primary colon tumor
• white blood cell (WBC) > 4000/cu.mm. and platelet count greater than or equal to 100,000/cu.mm.
• evidence of adequate renal (serum creatinine less than or equal to 1.5 mg%) and hepatic function (bilirubin less than or equal to 1.5 mg%; serum glutamic oxaloacetic transaminase (SGOT) less than or equal to 60 I.U./ml)
• performance status of 0, 1 or 2
• Patients with intestinal obstruction are eligible. Preliminary or complementary colostomy does not preclude entry of a patient provided randomization is carried out prior to the planned curative resection.
• The distal margin of the tumor must be greater than or equal to 12 cm from the anal verge as endoscopically measured with the patient in the knee-chest position.
• Carcinoembryonic antigen (CEA) must be performed pre-operatively but results need not be known at the time of randomization.

Exclusion criteria:

• malignant colon tumors other than carcinoma, i.e., sarcoma, lymphoma, etc.
• patients whose tumors demonstrate free perforation
• previous or concomitant malignancy, regardless of site - except patients with squamous or basal cell carcinoma of the skin, and carcinoma in situ of the cervix which have been adequately treated
• patients who have received prior treatment other than preliminary or complementary colostomy (radiation, chemotherapy or surgery) for their current malignancy.
• patients having tumors within 12 cm of the anal verge
• performance status of 3 or 4
• patients having non-malignant systemic disease (cardiovascular, renal, hepatic, etc.) which would preclude their being subject to the chemotherapy treatment option
• patients who are pregnant at the time of randomization
• patients with psychiatric or addictive disorders which would preclude obtaining informed consent
• patients who have multiple primary tumors involving both the colon and the rectum which would preclude them from being classified as having only colon cancer or only rectal cancer