Meta-Analysis of Defibrotide (DF) In the Treatment of Severe Hepatic Veno-Occlusive Disease (VOD) with Multi-Organ Failure (MOF) with Comparison to a Historical Control (HC).

Two prospective studies have investigated DF in the treatment of severe VOD/MOF post-SCT. Protocol 99–118 (a Phase 2 trial of DF; n=150) compared two doses of DF (25 and 40 mg/kg/day) and produced a similar complete response (CR) rate: 46% and 42% of patients (pts), respectively (Richardson et al BBMT 2010). Protocol 2005-01 (a Phase 3 trial enrolling pts with more severe VOD/MOF) demonstrated a CR of 24%. For the Phase 3 trial, a historical control (HC) had been created by review of 6864 medical charts at 35 BMT centers; up to 266 SCT pts were sequentially reviewed for symptoms of possible VOD/MOF with pts selected by an independent Medical Review Committee (MRC). The MRC was provided a redacted medical chart or pt narrative up to the date on which the pt met entry criteria, but remained blinded to outcome. Compared to this HC, the Phase 3 trial demonstrated that use of DF in pts with severe VOD/MOF resulted in a significantly higher CR rate (p=0.0148; Richardson et al ASH 2009). In order to assess consistency of DF treatment effect across these trials, a meta-analysis was undertaken.

Data for the subset of Phase 2 pts who had received a dose of 25 mg/kg/day and met eligibility criteria for the Phase 3 trial were combined with all 102 Phase 3 pts. Pts were required to have a diagnosis of VOD based upon Baltimore criteria (bilirubin ≥2.0 mg/dL with ≥2 of the following: hepatomegaly, ascites or 5% weight gain) with MOF as defined in the Phase 3 trial (renal and/or pulmonary dysfunction). CR was defined as bilirubin < 2 mg/dL + resolution of MOF. CR and mortality were assessed at D+100. DF (6.25 mg/kg IV q6h) was recommended for at least 21d. As the Phase 3 trial was not a prospectively randomized study, the primary efficacy analysis compared CR by D+100, adjusted by quintiles of propensity score based on 4 stratification variables of age, allogeneic SCT, prior SCT and ventilator/dialysis dependence at study entry.

This meta-analysis identified 31 Phase 2 pts treated at a dose of 25 mg/kg/day who met eligibility for the Phase 3 trial. Baseline characteristics were similar (n=31 and n=102, respectively); 61% vs 63% were male with median age (years) 20 vs. 22; acute leukemia in 52% and 44%; allogeneic SCT in 90% and 88%; and prior SCT in 9% and 13%. The 31 Phase 2 pts had less ventilator and dialysis dependence at study entry (6% and 3% vs 25% and 21%, respectively). For the HC, the MRC had selected 32 cases as having an unequivocal diagnosis of severe VOD, whose MOF was secondary to VOD and who met all entry criteria. DF treatment duration was shorter for the Phase 2 pts: a median of 15 vs 22 days of DF, respectively. Of the 133 pts, 38 (29%) achieved a CR by D+100, compared to 9% in the HC (p=0.0021 by propensity adjusted score). When comparing 133 pts with the HC, CR by D+100 rate was 43% for pediatric pts compared with 7% (p = 0.0001; n=58 and 14 pts, respectively). For pts with only 1 prior SCT and those with ≥2 prior SCTs, the CR rate was improved with DF (29% vs 10%, p=0.0008 with only 1 prior SCT; 25% vs 0%, p=0.0143 with 2+ prior SCTs). The CR rate was higher in pts not on dialysis/ventilator at study entry (35% vs 10%; p=0.0006). For those pts who were on dialysis/ventilator at study entry, the CR rate was 14% vs 8% (p=0.362). D+100 mortality rate equaled 60% vs 75% (p=0.0408 using Kaplan Meier estimates by log-rank analysis). Hemorrhage (any grade regardless of relationship to DF) was lower in the DF-group (61% vs 72% in the HC). Similar to the observation of decreased GvHD in a recent randomized prophylactic study (Corbacioglu et al, EBMT 2010), the incidence of GvHD (any grade) was lower (6% vs 25%) in the DF group.

This meta-analysis of DF-treated pts, pooling data from two separate clinical trials, confirmed improvement in CR and survival by D+100 compared to HC. The incidence of CR and D+100 survival was similar to that observed in prior phase 2 single-arm trials investigating DF in the treatment of this life-threatening condition. DF was well tolerated with less hemorrhage vs HC. A prospective sub-analysis suggests that DF results in a higher CR rate in pts who have not progressed to ventilator/dialysis dependence. GVHD was less common in DF treated pts compared to HC; this potentially important observation warrants further study.

Richardson:Gentium: Membership on an entity's Board of Directors or advisory committees. Grupp:Gentium: Membership on an entity's Board of Directors or advisory committees. Hume:Gentium: Employment. Massaro:Gentium: Consultancy. Hannah:Gentium: Consultancy. Iacobelli:Gentium: Employment.