High concentrations of tumor necrosis factor (TNF) alpha have been detected in the plasma of patients undergoing immunotherapy with interleukin 2 (IL-2), suggesting that this cytokine may play a role in the fever and shocklike state induced by the administration of high-dose IL-2. Dexamethasone has been shown to inhibit the synthesis of TNF by monocytes activated in vitro by endotoxin. To determine if dexamethasone can exert a similar suppressive effect on IL-2-induced TNF synthesis in vivo, the concentration of TNF alpha was measured in plasma samples serially obtained (a) from cancer patients participating in a phase I dose escalation clinical trial with high-dose IL-2 administered in conjunction with dexamethasone (IL-2/Dex) and (b) from patients participating in concurrent studies with IL-2 alone. In contrast to the high plasma levels of TNF alpha detected in patients receiving IL-2 alone, TNF levels in most of the IL-2/Dex patients remained below the threshold of detectability of our TNF radioimmunoassay. The concurrent administration of dexamethasone also prevented the IL-2-induced increase in serum levels of C-reactive protein, a hepatic acute phase reactant whose synthesis is regulated by proinflammatory cytokines such as TNF. The steroid-treated patients also failed to develop the neutrophil chemotactic defect characteristic of IL-2 recipients. The concomitant administration of dexamethasone increased the maximum tolerated dose of IL-2 approximately threefold and markedly reduced the hypotension and organ dysfunction ordinarily observed in these patients. These results demonstrate that dexamethasone inhibits the release of TNF into the circulation of patients undergoing immunotherapy with IL-2. They further suggest that the altered spectrum and reduced severity of IL-2 side effects observed in patients receiving dexamethasone may be attributable in part to the suppressive effect of steroids on IL-2-induced TNF synthesis.