Abstract 12310: A Randomized and Clinical Effectiveness Trial Comparing Two Pharmacogenetic Algorithms and Standard Care for Individualizing Warfarin Dosing: CoumaGen-II
BACKGROUND: Warfarin (W) therapy is characterized by marked individual dose variation and a narrow therapeutic window. Pharmacogenetic (PG) dosing of warfarin has been proposed to reduce variability and improve safety, but clinical trials evidence is meager.
METHODS: CoumaGen-II had 2 objectives (NCT00927862): 1) to compare (blinded, randomized) a 3-stage PG algorithm (PG3) with a previously tested 1-stage W initiation algorithm (PG1) (N=504, av. age 61y, 53% male), and 2) to compare PG-guided dosing with a standard empiric dosing protocol in a parallel control group (N=1866; av. age 57 y, 52% male). A rapid (1 h) method provided same-day genotyping for CYP2C9 and VKORC1. The primary outcome was % out-of-range (OOR) international normalized ratios (INRs) at 1 mo (1°) and 3 mo (2°). Primary analysis was modified intention-to-treat.
RESULTS: In the randomized comparison, PG3 was non-inferior to PG1 for the primary outcome at 1 and 3 mo but did not achieve superiority. However, the combined PG cohort was substantially superior to the parallel, empiric control cohort (%OOR INRs 31% vs 42% at 1 mo; 30% vs 42% at 3 mo, both p ≤ 1 x 10-9). Differences persisted after adjustment for age, sex, W indication, and inpatient versus outpatient initiation. PG-dosing also reduced adverse INR safety events (INR ≤ 1.5, ≥4.0) at 1 mo (0.92% vs 1.25%) and 3 mo (1.13% vs 1.87%), both p ≤ 1 x 10-5. First therapeutic INR tended to be achieved earlier with PG guidance (4.5 vs 4.9 d, p=0.08).
CONCLUSIONS: In a parallel-group, same-institution clinical effectiveness trial, PG-dosing of W substantially reduced %OOR INRs (35%-40% relative reductions) during the critical first 1-3 mo of W therapy, including safety events. More sophisticated PG-dosing (delaying CYP2C9 entry until day 3, using genetics to modify day 4-5 adjustment) added marginally (not significantly) to dosing-efficiency, suggesting that the simpler PG1 algorithm may be adequate for general clinical application. The substantial promise of PG-dosing in this clinical effectiveness trial awaits further validation in 3 ongoing randomized trials. :
- © 2011 by American Heart Association, Inc.