Summary. Severe hepatic veno-occlusive disease (VOD) is a recognized complication of autologous and allogeneic stem cell transplantation (SCT) that is often fatal. Defibrotide (DF) is a polydeoxyribonucleotide that has been found to have anti-thrombotic, anti-ischaemic and thrombolytic properties without causing significant anticoagulation. Preliminary studies have demonstrated activity for DF in the treatment of VOD, with minimal associated toxicity. In the present study, 40 patients who fulfilled established criteria for VOD were treated with DF on compassionate grounds in 19 European centres; 28 patients met risk criteria predicting progression of VOD and fatality or had evidence of multiorgan failure (MOF), and were defined as`poor-risk'. DF was commenced intravenously at a median of 14 d (range, 22 d to 53 d) post SCT at doses ranging from 10 to 40 mg/kg. The median duration of therapy was 18 d (range, 2±71 d). Twenty-two patients showed a complete response (CR) (bilirubin , 34´2 mmol/l and resolution of signs/symptoms of VOD and end-organ dysfunction) [CR 55%, confidence interval (CI) 40±70%] and 17 patients (43%) are alive beyond d 1100. Ten poor-risk patients showed a complete response (CR 36%, CI 21±51%). These results demonstrate that DF is an active treatment for VOD following SCT and a randomized trial is now underway in order to further evaluate its role.

Severe hepatic veno-occlusive disease (VOD) is a recognized complication of autologous and allogeneic stem cell transplantation (SCT) that is often fatal. Defibrotide (DF) is a polydeoxyribonucleotide that has been found to have anti-thrombotic, anti-ischaemic and thrombolytic properties without causing significant anticoagulation. Preliminary studies have demonstrated activity for DF in the treatment of VOD, with minimal associated toxicity. In the present study, 40 patients who fulfilled established criteria for VOD were treated with DF on compassionate grounds in 19 European centres; 28 patients met risk criteria predicting progression of VOD and fatality or had evidence of multiorgan failure (MOF), and were defined as 'poor-risk'. DF was commenced intravenously at a median of 14 d (range, -2 d to 53 d) post SCT at doses ranging from 10 to 40 mg/kg. The median duration of therapy was 18 d (range, 2--71 d). Twenty-two patients showed a complete response (CR) (bilirubin < 34.2 micromol/l and resolution of signs/symptoms of VOD and end-organ dysfunction) [CR = 55%, confidence interval (CI) 40--70%] and 17 patients (43%) are alive beyond d +100. Ten poor-risk patients showed a complete response (CR = 36%, CI 21--51%). These results demonstrate that DF is an active treatment for VOD following SCT and a randomized trial is now underway in order to further evaluate its role.

Raised PAI-1 levels post stem cell transplant are specific for VOD and a subsequent decrease in levels following treatment with DF may be associated with response to treatment.

The failure of the vaccine to produce a PSA response in the patients in the trial is not surprising considering the stage of the disease. The high PSA levels on entry indicate that the burden of disease was probably high and thus this was an extremely challenging group of patients in which to try and elicit a response through immunotherapy. However, the immunological evidence of a response to the vaccine was encouraging and suggests that further exploration of immunotherapy in less advanced disease may yield more encouraging clinical responses.

Gene therapy strategies based on modifying tumour cells using high efficiency adenoviral vectors have shown promise in the clinic. Recently the Coxsackie and adenovirus receptor (CAR) has been shown to mediate adenoviral entry into tumour cells, although previous studies also suggested a role for MHC class I heavy chain. Detailed evaluation of the expression of both CAR and MHC class I in prostate cancer cell lines would have important implications for therapeutic strategies. We have found that, unlike cell lines derived from other malignancies, in human and murine prostate cancer loss of CAR expression appears to be relatively infrequent and does not correlate with loss of MHC class I expression. These findings, together with the demonstration of appreciable levels of cellsurface expression of integrins, suggest that cancer vaccine strategies based on modifying whole prostate cancer cells should be feasible using the current generation of recombinant adenoviral vectors, without deleterious effects on either the virus vector or the target cell.