Efficacy of Recombinant Activated Factor VII (rFVIIa) vs. Activated Prothrombin Complex Concentrate (APCC) in Patients with Hemophilia with Inhibitors: A Bayesian Meta-Analysis.

INTRODUCTION: The recent FENOC comparative trial (Astermark et al. 2007) reported comparable efficacy for rFVIIa and APCC in the treatment of joint bleeds in patients with hemophilia complicated by inhibitors. A systematic review was combined with a Bayesian meta-regression analysis to place the FENOC results within the context of earlier non-comparative studies and to identify key variables influencing treatment efficacy.

METHODS: A systematic literature search was performed using electronic databases, web based trial registers, and grey literature. Clinical studies and/or trials in hemophilia patients with inhibitors receiving on-demand treatment with rFVIIa or APCC were included. Outcome of interest was the time of successfully stopping joint bleeds taking explicit account of repeated dosing. To model the latter, the regression model included a repeating Gompertz hazard function to compare the efficacy of standard rFVIIa treatment (one 90 μg/kg rFVIIa infusion every 3 hrs) with standard APCC treatment (one 75 IU/kg APCC infusion every 12 hrs). The hazard function (the conditional probability of ending the bleeding) was characterized by a repeated pattern of fixed increases after each infusion and by subsequent decreases thereafter. It was assumed that each sequential infusion had the same effect on the hazard resulting in a shark tooth like function. Covariates included were medication type, medication type combined with dosage (measured in μg/kg for rFVIIa and IU/kg for APCC), efficacy rating scale method, and efficacy rater.

RESULTS: Of the 17 studies included in the systematic review, 14 provided adequate data to be included in the Bayesian model, reporting on almost 2,000 joint bleeds combined (of which 96 were from the FENOC itself). The analysis demonstrated that medication type combined with dosage appears to have a statistically significant effect on efficacy of treatment. For rFVIIa the model estimated that 75%, 94%, and 98% of bleeds would be treated effectively after 12, 24, and 36 hours respectively. In contrast, for APCC the model estimated that 39%, 63%, and 78% of bleeds would be treated effectively after 12, 24 and 36 hours respectively.

CONCLUSIONS: The FENOC study demonstrated similar efficacy levels for treatment with rFVIIa and APCC. In contrast, the present meta-analysis including 14 studies and almost 2,000 joint bleeds estimated that standard treatment with rFVIIa is likely to result in higher efficacy levels at the 12, 24 and 36 hour time points than treatment with APCC. This implies that treatment with rFVIIa may be associated with a faster time to joint bleed resolution than APCC.