Genetics-Based Dosing Strategies

Brian F. Gage, MD, MSc, a professor of medicine at Washington University School of Medicine in St. Louis, recently completed the Genetics-Informatics Trial (GIFT) of Warfarin Therapy to Prevent Deep Venous Thrombosis.

The study tested the effectiveness and safety of genetically guided warfarin dosing in nearly 1600 older patients who started on the therapy to limit the risks of bleeding and clotting after hip or knee replacement surgery. Patients who received genetically guided dosing experienced fewer adverse events and were less likely to overdose on warfarin than patients who received clinical dosing, according to the study.

In an interview, Dr Gage said the genes that regulate warfarin metabolism and sensitivity and vitamin K use are highly variable, so providers can't simply look at patients and predict their therapeutic warfarin dose. That guesswork could be avoided, thanks to the study’s findings, which Dr Gage said showed promise for improving on the trial-and-error approach of standard dosing methods.

What led you to explore the genetics-based dosing strategy for warfarin?

For two decades, I ran an anticoagulation clinic with clinical pharmacists. While managing warfarin therapy, we noticed that some elderly or petite patients required high warfarin doses, while other patients overdosed on small doses. Before the GIFT trial, we identified novel single nucleotide polymorphisms that correlated with the therapeutic warfarin dose. We had a good idea of how genes and clinical factors affected the dose of warfarin, but we didn't know whether genotype testing improved clinical outcomes. In GIFT, we tested the hypothesis that prospective genotyping increased warfarin safety and found that pharmacogenetic testing improved INR control and reduced overall adverse events by 27%.

How did your current study differ from other research on this topic and what were the important findings?

Prior multicentered trials—Clarification of Optimal Anticoagulation Through Genetics (COAG) and European Pharmacogenetics of Anticoagulant Therapy (EU-PACT)—reached contradictory results about the pharmacogenetic dosing of warfarin. In EU-PACT, genetic-based dosing improved INR control and allowed patients to reach therapeutic INR more quickly than a fixed starting warfarin dose. In COAG, there was no benefit of genetic-based dosing compared with optimized clinical dosing. Like COAG, GIFT was a double-blind comparison of randomized genetic-based dosing and clinical dosing optimized with a clinical algorithm. Unlike COAG and EU-PACT, however, GIFT included an additional gene—CYP4F2—and genetic-dosing algorithms through day 11 of therapy. Also, GIFT was larger than those two trials combined.

What are the limitations of this dosing strategy and how could they be solved?

There are a couple limitations: access to genetic dosing algorithms and reimbursement for genotyping. To address the need for access, we provide dosing algorithms through a non-profit website (www.WarfarinDosing.org). Maintaining this site is expensive and time-consuming—it currently includes algorithms for day 1 through 5 of therapy, but we need to add algorithms for day 6 through 11. The National Heart, Lung, And Blood Institute (NHLBI) branch of the National Institutes of Health (NIH) currently backs the website, but it would benefit from foundational support. The algorithms could be more easily accessed if they were integrated into electronic medical records.

The second limitation to warfarin genotyping is reimbursement. The Centers for Medicare and Medicaid Services (CMS), along with the NHLBI, supported the GIFT trial. CMS will use the study’s results to decide whether it should cover the cost of genotyping. If CMS decides to fund testing for Medicare patients, large medical centers will be able to offer pharmacogentic tests for patients who are starting warfarin therapy and results could be available within a day.

How would genetic-based dosing work in clinical practice? Would it help solve some of the practical drawbacks of monitoring INR levels?

We can use algorithms to estimate the therapeutic dose as functions of age, height, drug interaction, and genotype. There’s a different algorithm for days 1 through 11 of therapy, and those algorithms predict the final therapeutic dose as a function of those clinical and genetic factors. Including the algorithms in electronic medical records might be an efficient way to apply genetic dosing. Electronic medical records could also export patient information to warfarindosing.org, so that clinicians don’t have to enter it manually. Importantly, genetic-based dosing can improve the safety of warfarin initiation, but it doesn’t obviate the need for regular INR testing and dose management.

What are the big-picture implications of your findings?

Pharmacogenetic dosing worked! We’ve seen precision medicine prove effective for some chemotherapy, but, after COAG, we were becoming skeptical of using that approach for cardiovascular drugs. GIFT found that genotype-based dosing is effective for warfarin and suggested the approach deserves further study. Although new oral anticoagulants (NOACs) have become popular for some indications, warfarin is still commonly used. Also, warfarin is cheap and easy to reverse. If we can decrease the risk of using the drug, we should be able to prevent adverse events for a lot of patients. Using genetic-based management of warfarin dosing means patients are less likely to overdose and experience complications. The approach is particularly important in older patients, because warfarin is the medication that’s most likely to land them in emergency room for treatment of therapy-related adverse events.

—Dan Cook