Systemic release of soluble TNF receptors after high-dose TNF in isolated limb perfusion

Isolated limb perfusion (ILP) with high-dose tumour necrosis factor (TNF), interferon γ and melphalan (TIM) is an efficient treatment for patients with regionally advanced melanoma and sarcoma. In 44 patients, we determined the kinetics of soluble TNF receptors (sTNF-RI and RII) plasma concentrations, and correlated them with systemic TNF and interleukin 6 (IL-6) levels and shock. Seven patients treated conventionally by ILP without cytokine served as controls. Elevated levels of both sTNF-Rs were observed within 30 min after begining of the TIM-ILP. A first peak of sTNF-Rs levels was observed 3 h after ILP and was followed by a rapid decrease reaching a nadir at 12-14 h post ILP. This first peak was followed by a second, long-lasting elevation of both sTNF-Rs levels persisting for 4 to 5 days after TIM-ILP. Patients treated by ILP without TNF/interferon γ (IFN-γ) had no detectable increase in either sTNF-Rs or in circulating TNF, demonstrating that the release of TNF-Rs was dependent upon the administration of TNF/IFN-γ. High plasma levels of TNF and IL-6 were observed in patients that had more than 5% leakage during the TIM-ILP, but no significant correlation between TNF levels and the peak values of both sTNF-Rs was observed. The levels of TNF and IL-6 were, however, significantly related to each other. TNF systemic levels, but not sTNF-Rs concentrations, correlated significantly with the severity of the shock observed after TIM-ILP. Patients in which sTNF-RII concentration was in excess over circulating TNF, had no shock or grade I shock only, suggesting that sTNF-RII may play a protective, although limited, role in inhibiting activity of circulating TNF.