We have previously shown in rats that the provision of analgesic doses of morphine significantly reduces the tumor-promoting effects of undergoing and recovering from surgery. Because morphine had no effect in non-operated animals, and because a single preoperative dose given hours before tumor inoculation was effective, we have suggested that it is the pain-relieving effects of the drug that underlies its beneficial impact. To support and strengthen this suggestion, two different regimens of analgesia were employed, the systemic administration of the more selective mu-agonist, fentanyl, and the intrathecal (i.t.) administration of bupivacaine plus morphine. To assess host resistance against metastasis, we used a lung clearance assay of the MADB106 mammary adenocarcinoma, a natural killer (NK)-sensitive syngeneic cell line that metastasizes only to the lungs. Female and male Fischer 344 rats were randomly assigned to one of four groups using a 2x2 experimental design: experimental laparotomy under halothane anesthesia versus anesthesia alone, by drug treatment versus vehicle. In the first in vivo experiment, fentanyl was administered 20 min before surgery (40 microg/kg subcutaneously (s.c.)), and at the end of surgery in a slow-release suspension (20 microg/kg s.c.). In the second in vivo experiment, bupivacaine (10 microg) plus morphine (20 microg) in 50 microl was administered i.t. before surgery. Surgery resulted in a 3- to 4-fold increase in the lung retention of MADB106 cells in both males and females, and the observed surgery-induced increase in lung tumor retention was reduced by more than 65% in the fentanyl-treated animals and more than 45% in the animals receiving i.t. bupivacaine plus morphine. Neither drug regimen exerted effects in the anesthesia only animals. Surgery also resulted in a significant suppression of whole blood NK activity assessed at 5 h postoperatively, the same time point at which MADB106 tumor cells were inoculated in the in vivo studies. Unlike the in vivo study, fentanyl suppressed NK activity at this time point in non-operated rats, but had no effect in operated rats. Taken together, these findings strengthen the suggestion that the management of perioperative pain is a critical factor in preventing surgery-induced decreases in host resistance against metastasis. If similar relationships between pain and metastasis occur in humans, then pain control must become a priority in the postoperative care of individuals with cancer.