Table 1.
Association of pharmacogenomics with warfarin outcomes.
| Studies | Design | N | Population | Alleles | Outcomes | P |
|---|---|---|---|---|---|---|
| Positive association | ||||||
| Primohamed et al., 2013 [32] | RCT, genotype guided vs. standard dose | 455 | 98% White 1% Black 1% Asian |
CYP2C9*2 CYP2C9*3 VKORC1*2 |
Improved time within therapeutic INR (67.4% vs. 60.3%); reduction in INR > 4, reduced time to therapeutic INR | <0.001 |
| Gage et al., 2017 [37] | RCT, genotype guided vs. clinical | 1597 | 91% White 6% Black 2% Asian 1% Other |
CYP2C9*2 CYP2C9*3 VKORC1*2 CYP4F2*3 |
Reduced composite measure of major bleeding, INR > 4, death, and VTE (10.8% vs. 14.7%). In hip and knee arthroplasty patients |
<0.02 |
| Caraco et al., 2008 [38] | RCT, Genotype vs. clinical | 191 | Unavailable |
CYP2C9*2 CYP2C9*3 |
Reduction in time to first therapeutic INR (2.73 days earlier) and reduction in time to stable INR (18.1 days earlier) | <0.001 |
| Gage et al., 2008 [28] | Validation of dosing algorithm | 292 | 93% Caucasian 15% Black 2% Hispanic |
CYP2C9*2 CYP2C9*3 VKORC1*2 |
Pharmacogenomic dose prediction more accurate than clinical dose prediction (53% vs. 17% of explained variability, respectively) | <0.0001 |
| IWPC, 2009 [29] | Validation of dosing algorithm | 1009 | 55% White 30% Asian 10% Black 5% Other |
CYP2C9*2 CYP2C9*3 VKORC1*2 # |
Pharmacogenomic dose prediction more accurate than clinical dose prediction (accurately identified 49.4% vs. 33.3% of patients requiring ≤21 mg warfarin per week, respectively) | <0.001 |
| Gong et al., 2011 [31] | Validation of dosing algorithm | 167 | 95% White 2% Black 2%Asian 1% Other |
CYP2C9*2 CYP2C9*3 VKORC1*2 CYP4F2*3 |
Demonstrated the safe effective prediction of dose limiting variation | N/A |
| Negative association | ||||||
| Kimmel et al., 2013 [33] | RCT, Genotype guided vs. clinical | 1015 | 66%White 27% Black 7% Hispanic |
CYP2C9*2 CYP2C9*3 VKORC1*2 |
No difference in time in therapeutic INR (45.2% vs. 45.4%) No difference in anticoagulation control or dose prediction |
0.91 |
| Verhoef et al., 2013 [39] | RCT, Genotype guided vs. clinical | 1597 | 98% White |
CYP2C9*2 CYP2C9*3 VKORC1*2 |
No difference in time in therapeutic INR range (61.6% vs. 60.2%) | 0.47 |
| Pengo et al., 2015 [34] | RCT, Genotype guided vs. standard | 180 | 100% White |
CYP2C9*2 CYP2C9*3 VKORC1*2 CYP4F2*3 |
No difference in out of range INRs (45.6% vs. 43.6%) or time in therapeutic INR range (51.9% vs. 53.3%) | 0.79 0.71 |
| Anderson et al., 2007 [40] | RCT, Genotype guided vs. standard | 200 | 94% White |
CYP2C9*2 CYP2C9*3 VKORC1*2 |
No difference in time in therapeutic INR range (30.7% vs. 33.1%) | 0.47 |
RCT, randomized control trial; INR, International normalization ratio; VTE, venous thromboembolism; #, VKORC1*2 or one of six other linked SNPs, N/A, not available.