Meta-analysis of randomized controlled trials reveals an improved clinical outcome of using genotype plus clinical algorithm for warfarin dosing.

Previous studies have raised interest in using the genotyping of CYP2C9 and VKORC1 to guide warfarin dosing. However, there is lack of solid evidence to prove that genotype plus clinical algorithm provides improved clinical outcomes than the single clinical algorithm. The results of recent reported clinical trials are paradoxical and needs to be systematically evaluated. In this study, we aim to assess whether genotype plus clinical algorithm of warfarin is superior to the single clinical algorithm through a meta-analysis of randomized controlled trials (RCTs). All relevant studies from PubMed and reference lists from Jan 1, 1995 to Jan 13, 2014 were extracted and screened. Eligible studies included randomized trials that compared clinical plus pharmacogenetic algorithms group to single clinical algorithm group using adult (≥ 18 years) patients with disease conditions that require warfarin use. We further used fix-effect models to calculate the mean difference or the risk ratios (RRs) and 95% CIs to analyze the extracted data. The statistical heterogeneity was calculated using I(2). The percentage of time within the therapeutic INR range was considered to be the primary clinical outcome. The initial search strategy identified 50 citations and 7 trials were eligible. These seven trials included 1,910 participants, including 960 patients who received genotype plus clinical algorithm of warfarin dosing and 950 patients who received clinical algorithm only. We discovered that the percentage of time within the therapeutic INR range of the genotype-guided group was improved compared with the standard group in the RCTs when the initial standard dose was fixed (95% CI 0.09-0.40; I(2) = 47.8%). However, for the studies using non-fixed initial doses, the genotype-guided group failed to exhibit statistically significant outcome compared to the standard group. No significant difference was observed in the incidences of adverse events (RR 0.94, 95% CI 0.84-1.04; I(2) = 0%, p = 0.647) and death rates (RR 1.36, 95% CI 0.46-4.05; I(2) = 10.4%, p = 0.328) between the two groups. Allocation to genotype plus clinical algorithm may be associated with a significant improvement of the percentage of time within the therapeutic INR range for patients adopting fixed dose of warfarin. The incidence of total adverse events and death rates did not differ between these two groups. Further experiments need to be conducted to confirm these findings.