Defibrotide for the Treatment of Hepatic Veno-Occlusive Disease/Sinusoidal Obstruction Syndrome with Multi-Organ Dysfunction: Final Results from a Pivotal, Historically Controlled, Phase 3 Trial

Introduction

Hepatic veno-occlusive disease/sinusoidal obstruction syndrome (VOD/SOS), is a rare and potentially life-threatening complication of hematopoietic stem cell transplantation (HSCT). Severe cases, historically defined by multi-organ dysfunction (MOD), may be associated with mortality rates of >80%. There is no FDA-approved treatment for VOD/SOS. Defibrotide (DF) has a proposed mechanism of action that includes stabilization of endothelial cells and restoration of thrombo-fibrinolytic balance.

Earlier analyses of a pivotal phase 3 trial of DF in VOD/SOS plus MOD (Richardson et al. Blood. 2009;114:Abstract 654) underpinned approval of DF in the EU to treat severe hepatic VOD/SOS after HSCT. Additional data were obtained at the request of US health authorities. Here we present the final analysis: day +100 survival (primary endpoint) and complete response (CR; secondary).

Methods

This was a multicenter, open-label, phase 3 historical control (HC) study assessing DF. Eligible patients met Baltimore VOD/SOS criteria (total bilirubin ≥2.0 mg/dL with ≥2 of: hepatomegaly, ascites, or 5% weight gain) by day +21 post-HSCT, plus MOD (renal [trebling of creatinine levels, reduced creatinine clearance, or dialysis] and/or pulmonary [oxygen saturation ≤90%, need for oxygen supplementation/ventilator dependence]) by day +28 post-HSCT. Exclusion criteria included severe graft-versus-host disease (GvHD) of liver or gut, clinically significant bleeding, or need for ≥2 pressors. HC patients were reviewed for inclusion/exclusion criteria in a sequential review of medical charts starting 6 months prior to use of DF at each site; a blinded medical review committee made the final determination of HCs unequivocally meeting criteria for VOD/SOS with MOD. DF dose was 25 mg/kg/d in 4 divided 2-hour IV infusions q6h; recommended treatment duration was ≥21 days. Primary endpoint was day +100 survival. CR by day +100 was a secondary endpoint. Treatment difference in survival and CR rates and their 95% confidence intervals were estimated using propensity-stratified and weighted (Koch-adjusted) estimates of differences in proportions that account for baseline prognostic factors of survival (ie, ventilator and/or dialysis dependency at entry, age ≤/>16 years, transplant type, and prior HSCT). Analyses included patients treated with DF and HCs.

Results

There were 102 patients in the DF group and 32 cases selected as HCs. Baseline characteristics were similar in the DF and HC groups: mean age (26 and 25 years; 43% and 44% ≤16 years), allogeneic graft (88% and 84%), prior HSCT (13% and 9%), ventilator- and/or dialysis-dependent at study entry (33% and 22%), myeloablative conditioning (87% and 94%), and the most common underlying diseases (acute leukemias: 45% and 47%), respectively. In the DF-treated group, common GvHD medications included tacrolimus (49%), methotrexate (41%), and cyclosporine (38%); in the HC group, common medications were cyclosporine (72%) and methotrexate (63%).

Survival at day +100 in the DF and HC groups was 38% and 25%, respectively. The propensity-stratified difference in survival was 23.0% (95.1% CI, 5.2-40.8, P = .0109). Respective observed CR rates by day +100 were 25.5% and 12.5%, and the propensity-stratified difference in CR was 19.0% (95.1% CI, 3.5-34.6, P = .0160). Comparing the earlier EU and final analyses, the survival rates at day +100 in each group did not vary; however, the propensity adjusted final analysis provided a different level of statistical significance. Day +100 CR rates in the original analysis were slightly lower in both arms at 24% and 9% due to increased data capture to investigate CR; the P value was essentially unchanged. For the DF group, 45% had an adverse event (AE) at least possibly related to study drug, and 21% had a serious AE at least possibly related to study drug. In this very sick population, percentages of patients with ≥1 AE leading to death were similar between DF and HC patients (64% and 69%), as were hemorrhagic AEs (64%, 75%) and hypotension (39%, 50%).

Conclusions

Based on observed study data and using a propensity-adjusted rate difference estimator, patients treated with DF had a 23% reduction in risk of death by day +100 and 19% improvement in CR rate. Overall incidence of hemorrhage and fatal AEs were similar between groups with AEs consistent with those expected in this critically ill population.

Support: Jazz Pharmaceuticals.