Meta‐analysis of observational studies of full‐length and B‐domain‐deleted factor VIII for prophylaxis – a standard statistical approach

Abstract

Summary. Meta‐analyses of observational studies have become increasingly common to support evidence‐based clinical decisions. We analyzed currently available clinical and pharmacokinetic studies of full‐length factor VIII (FL‐FVIII) vs B‐domain deleted recombinant factor VIII (BDD‐rFVIII) to investigate possible differences in clinical efficacy. A standard approach to observational study meta‐analysis was used as recommended by the Meta‐analysis Of Observational Studies in Epidemiology (MOOSE) consensus group. Data from included studies were combined using a random effects model to accommodate between‐study heterogeneity. A multivariate Poisson regression analysis was used to calculate the incidence rate ratio for bleeding. Justification for the appropriate use of this statistical approach is given. The meta‐analysis suggested reduced haemostatic efficacy during prophylaxis and shortened half‐life of BDD‐rFVIII as compared with FL‐FVIII. This analysis was rigoroulsy conducted using well‐accepted statistical methodology and presents the most comprehensive evidence‐based assessment currently available on risk factors for breakthrough bleeding under factor VIII prophylaxis. Our conclusions have been strengthened by subsequently introduced prescribing information and regulatory action for BDD‐rFVIII.