Systemic and haemodynamic toxicity after isolated limb perfusion (ILP) with TNF-α

The aim of this study was to evaluate the systemic and haemodynamic postoperative effects of ILP with medium-low dose of TNF alpha in patients diagnosed with primary or recurrent limb melanoma or sarcoma, and to compare the resulting toxicity with Systemic Inflammatory Response Syndrome (SIRS). A prospective study on 17 consecutive patients with primary or recurrent limb tumor (melanoma or sarcoma) subjected to ILP with escalating doses of TNF alpha (0.5-2.0mg) was carried out. Seventeen patients with primary or recurrent limb melanoma or sarcoma were subjected to ILP with escalating doses of TNF alpha. ILP was carried out with the standard techniques, blood being warmed at 42°C for an hour. Serial serum TNF alpha determinations were performed before, during and after limb perfusion in nine patients. Systemic and pulmonary haemodynamics, by a radial and pulmonary artery catheter inserted before the induction of anesthesia, were monitored at 5 different times: before the induction of anesthesia (TO), and 6, 12, 24 and 48 hours after treatment (T1-4). Complete isolation of the limb was not always achieved, therefore leakage of TNF alpha occurred frequently during the perfusion in all patients with maximum systemic TNF alpha concentrations ranging from 431 to 111000 pg/ml. After perfusion only two patients showed detectable TNF alpha levels in peripheral blood which returned to baseline values within nine hours. These two patients had serious systemic toxicity: shock and respiratory failure secondary to pulmonary edema. Acute pulmonary edema was also observed in another patient. All three cases required supportive therapy provided by means of mechanical ventilation. In the remaining 14 patients a sepsis-like syndrome was observed. The most significant haemodynamic changes were due to the CO, which rose by 35%, and the SVR, which remained consistently low throughout. A reduction in Hb was observed in all patients (with an average decrease of 4 g/dl), while DO 2 and VO 2 levels rose, though not to statistically significant levels. Hypoxia occurred in all 14 patients. In three of the remaining 14 cases bilateral pulmonary leaks were noted, however the use of mechanical ventilation was not required. No perioperative death occurred and the aforementioned side effects were all reversible resulting in a patient's mean postoperative ICU permanence of 4 days (range 3 to 7 days). In conclusion, ILP with TNF alpha induces cardiovascular, respiratory and hematological toxicity with haemodynamic parameters being similar to those noted in SIRS probably due to leakage of TNF alpha in the systemic circulation during the perfusion. Nevertheless, this systemic toxicity was short-lived resulting in an acute reaction following a single application.