Pharmacogenetics-based Warfarin Dosing Algorithm Decreases Time to Stable Anticoagulation and the Risk of Major Hemorrhage

Abstract

Warfarin is yet the most widely used oral anticoagulant for thromboembolic diseases, despite the recently emerged novel anticoagulants. However, difficulty in maintaining stable dose within the therapeutic range and subsequent serious adverse effects markedly limited its use in clinical practice. Pharmacogenetics-based warfarin dosing algorithm is a recently emerged strategy to predict the initial and maintaining dose of warfarin. However, whether this algorithm is superior over conventional clinically guided dosing algorithm remains controversial. We made a comparison of pharmacogenetics-based versus clinically guided dosing algorithm by an updated meta-analysis. We searched OVID MEDLINE, EMBASE, and the Cochrane Library for relevant citations. The primary outcome was the percentage of time in therapeutic range. The secondary outcomes were time to stable therapeutic dose and the risks of adverse events including all-cause mortality, thromboembolic events, total bleedings, and major bleedings. Eleven randomized controlled trials with 2639 participants were included. Our pooled estimates indicated that pharmacogenetics-based dosing algorithm did not improve percentage of time in therapeutic range [weighted mean difference, 4.26; 95% confidence interval (CI), -0.50 to 9.01; P = 0.08], but it significantly shortened the time to stable therapeutic dose (weighted mean difference, -8.67; 95% CI, -11.86 to -5.49; P < 0.00001). Additionally, pharmacogenetics-based algorithm significantly reduced the risk of major bleedings (odds ratio, 0.48; 95% CI, 0.23 to 0.98; P = 0.04), but it did not reduce the risks of all-cause mortality, total bleedings, or thromboembolic events. Our results suggest that pharmacogenetics-based warfarin dosing algorithm significantly improves the efficiency of International Normalized Ratio correction and reduces the risk of major hemorrhage.

Genotype-guided dosing versus conventional dosing of warfarin: A meta-analysis of 15 randomized controlled trials

“…We also meta‐analysed other endpoints about adverse events, including thromboembolism and all‐cause mortality. However, our results did not demonstrate superiority of reduced risks of these adverse events in GD group as compared with CD group, which were in consistent with previous meta‐analyses …”

“…Results demonstrated that GD algorithm significantly shortened time to first therapeutic INR and time to reach stable therapeutic dose, and reduced the risk of warfarin‐related major bleedings. GD algorithm presented similar PTTR at ≤1 month and higher PTTR at >1 month follow‐up compared to CD method, which is in accordance with the finding of previous meta‐analyses . There was no significant difference in the incidence of patients achieving stable dose between GD and CD groups at ≤1 month follow‐up, but more patients in the GD group achieving stable dose at >1 month follow‐up.…”

“…However, GD was associated with approximate 23% less incidence of INR >4 compared fixed‐dose regimen, but also failed to demonstrate a significant difference. The result was identical to that of previous meta‐analyses …”

“…GD algorithm presented similar PTTR at ≤1 month and higher PTTR at >1 month follow-up compared to CD method, which is in accordance with the finding of previous metaanalyses. 13,16 There was no significant difference in the incidence of patients achieving stable dose between GD and CD groups at ≤1 month follow-up, but more patients in the GD group achieving stable dose at >1 month follow-up. However, the risks of INR >4, all bleeding, thromboembolism and all-cause mortality were similar between GD vs CD groups.…”

“…Moreover, a few mistakes were made in some of the meta‐analyses. For instance, Shi et al and Wang et al extracted incorrect data for the endpoint of time to stable dose, and Tang et al categorized the control group of one study as fixed‐dose regimen by mistake. The latest meta‐analyses included 2 trials grouping patients by pseudo‐random, which could bring about great bias.…”