Skip to main content
Premium Trial:

Request an Annual Quote

Study Shows PGx-based Warfarin Dosing Algorithm Beats INR Alone, Even in Second Week of Treatment


By Molika Ashford

A multi-institution research team led by Intermountain Medical Center in Utah has created an algorithm combining clinical and genomic information that could inform more accurate warfarin dosing as patients move into their second week after initiation of the anti-coagulant therapy.

The new nomogram, which the team described last month in the journal Thrombosis and Haemostasis, follows on earlier pharmacogenetic algorithms for finding a best starting dose, as well as equations to guide treatment into the first few days after initiation. The group intends to integrate the new work with these earlier guides as part of the application, created by Brian Gage, a researcher at Washington University in St. Louis and a senior author of the study.

Warfarin dosing has been likened to a tightrope walk, as wavering to either too much or too little by even a small percentage can lead to fatal bleeding or blood clots.

In its report on the new algorithm, the Intermountain team wrote that even in the second week of treatment — when physicians can rely on measurements of patient’s international normalized ratio, or INR, response to guide dosing — a strategy including both genotype and INR did a better job predicting correct warfarin doses than a predictor using only INR and other clinical factors. The PGx algorithm could explain at least 70 percent of inter-patient variability, while an algorithm without genomic factors could explain only about 60 percent, the group found.

Benjamin Horne, the study’s first author and director of genetic epidemiology at Intermountain's heart institute told PGx Reporter that the current paper extends the group’s earlier work to determine how long genetic factors remain relevant in choosing correct warfarin doses.

"We wanted to know if it was still relevant after a week, or if INR told you everything you needed to know," he said. "We found that yes, even when you have two INR measurements after a week of therapy, genetics still can contribute."

Gage, who spurred the group's formation and earlier work on PGx-guided dosing in the first few days of treatment, wrote in an e-mail that the team’s hope is for the new algorithm to "allow clinicians to maximize the benefit we get from genotype."

Demonstrating that genomic information continues to be relevant to dosing after initiation is important, the study authors wrote, because current genotyping turnaround times "limit the ability to initiate dosing based on PGx information."

The FDA first changed warfarin's label in 2007 to reflect the influence of the CYP2C9 and VKORC1 genes in metabolizing the drug, then updated the label in 2010 to include specific pharmacogenomically guided dosing ranges (PGx Reporter 2/3/2010).

Despite these labeling changes, however, pharmacogenetic testing to dose warfarin has not been adopted by a majority of US healthcare providers mainly due to the fact that physicians are comfortable using standard INR measurement. Moreover, healthcare providers and payors largely believe that the turnaround time for genetic test results is too long to allow PGx-guided warfarin dosing to be used in acute settings where there is a pressing need for quick and accurate dosing (PGx Reporter 12/14/2011).

Horne and his team are hoping to change this perception since the study results suggest that even if patients are initiated into therapy agnostically, genotype can still help improve their dosing later in treatment.

Furthermore, Horne said the ability to predict dosage using these algorithms actually increases as the therapy progresses. "The algorithms to predict what people should start on are pretty good. Not as good as this last one, but for the amount of information you have at initiation, they do a pretty good job: you can predict about 50 percent of the variation in [what a stable dose should be]."

"But after seven days, you can predict more, about 70 percent," he said, as more clinical information accrues.

In their latest study, the team tracked a total of 1,342 patients on days six through 11 of their warfarin treatment to derive a formula combining genotype and physical and clinical factors, like weight, gender, and INR, to help guide correct dosing.

Genotyping for CYP2C9*2 and *3 was available for all subjects and some centers also measured *5 and *6 mutations, which the group included "to improve accuracy in African ancestry populations." The researchers also obtained genotype for VKORC1-1639 G>A mutations, or where that genotype was not available, inferred type by measuring other genes in high linkage disequilibrium, the authors wrote.

The team created one algorithm combining only clinical factors, and another that added genotype. The researchers then applied these to an independent retrospective group of 342 patients to see how well the two could predict patient response to warfarin. They also tested a small prospective cohort of 43 orthopedic patients to confirm the safety and accuracy of the clinical-only algorithm.

Overall, the clinical algorithm explained about 61 percent of the variation in therapeutic dose, while the PGx algorithm explained about 70 percent. For both algorithms, the accuracy increased with advancing days of treatment.

"You can get to the appropriate dose by trial and error, supplemented by INR right now," Horne said. "But we found that using genetics, it contributes information you wouldn’t otherwise obtain even after seven days."

What is still unclear from this study and others is whether PGx-guided dosing actually improves patients' outcomes on the drug. "That’s the next question," Horne said.

The group’s new PGx algorithm is now being investigated as part of Gage’s Genetics Informatics Trial of Warfarin, or GIFT, trial.

Small studies have investigated the overall benefit of PGx warfarin initiation algorithms, but found no significant improvement in patient outcomes. In GIFT the researchers hope to show that the demonstrated validity of these algorithms used past the first dose might translate into real clinical benefit.

Private payors do not currently reimburse for genetic testing to dose warfarin due to limited clinical utility data showing that testing improves patient outcomes and saves money compared to standard INR monitoring. The Centers for Medicare & Medicaid Services agreed to pay for testing only in the context of a clinical trial that shows such an intervention improves outcomes for Medicare beneficiaries (PGx Reporter 3/9/2011).

Gage said the GIFT trial is ongoing and blinded, so there is no indication yet as to whether it may reveal an association between PGx strategies and better patient outcomes enough to justify the costs of genotyping. "We don’t know if genotyping has prevented adverse events, but that is our hypothesis," he said.

Gage said the plan is to now also incorporate the group's week-two algorithm into, which already uses the researchers' previous models for the first days of treatment.

He said the site is averaging about 2,000 visitors per week.

Have topics you'd like to see covered in PGx Reporter? Contact the editor at mashford [at] genomeweb [.] com.

The Scan

Y Chromosome Study Reveals Details on Timing of Human Settlement in Americas

A Y chromosome-based analysis suggests South America may have first been settled more than 18,000 years ago, according to a new PLOS One study.

New Insights Into TP53-Driven Cancer

Researchers examine in Nature how TP53 mutations arise and spark tumor development.

Mapping Single-Cell Genomic, Transcriptomic Landscapes of Colorectal Cancer

In Genome Medicine, researchers present a map of single-cell genomic and transcriptomic landscapes of primary and metastatic colorectal cancer.

Expanded Genetic Testing Uncovers Hereditary Cancer Risk in Significant Subset of Cancer Patients

In Genome Medicine, researchers found pathogenic or likely pathogenic hereditary cancer risk variants in close to 17 percent of the 17,523 patients profiled with expanded germline genetic testing.