Eleven hundred and sixteen primary female breast cancer patients received one short perioperative chemotherapy course. The node negative patients were randomised between immunotherapy (corynebacterium parvum s.c. around the scar 2 weeks after mastectomy), or no further adjuvant therapy. A moderate, but significant delaying effect was observed, without side effects. The node positive patients were randomised to four groups: 1) the same immunotherapy as in the node negative patients, 2) CMF for 1 year, 3) combination of these two treatments, or 4) no further adjuvant therapy. The prolonged chemotherapy had a significant positive effect, but also considerable and distressing side effects. The immunotherapy had a non-significant negative effect in the node positive patients, but without side effects.