Patients' genetic profiles can help avoid incorrect dosing of a common, dangerous drug, say Stanford scientists

- By Krista Conger

STANFORD, Calif. - Your DNA affects your hair color, your blood type and even how you react to some medications. Now scientists have taken a big step toward folding your personal genetic profile into many of the prescriptions you carry away from the pharmacy. Most immediately, the advance will likely lead to the safer, more effective use of a common anticoagulant called warfarin that, when taken in the wrong amount, can be very dangerous. In the future, it could affect how doctors prescribe dozens of common medications.

Because an effective amount of warfarin can vary by as much as 10-fold among individuals, physicians currently start most of their patients on a low dose and ramp up gradually until blood tests indicate the dose is correct - a process that can take months.

"It appears that up to 46 percent of people will require a warfarin dose that is significantly higher or lower than average," said Russ Altman, MD, PhD, professor of bioengineering, genetics and medicine at Stanford University School of Medicine, noting that about 800,000 people are likely to be affected in the United States alone. "We're hoping that our research will help clinicians get it right on the first try."

Altman, who chairs Stanford's bioengineering department, and Teri Klein, PhD, a senior research scientist in genetics at the School of Medicine, solicited, in collaboration with others, warfarin dose-response information from over 21 sites - known as the International Warfarin Pharmacogenetics Consortium - in nine countries for inclusion in the study, which was published in the Feb. 19 issue of the New England Journal of Medicine.

Although this study is not the first to show the advantage of incorporating genetic information into prescribing patterns, it is by far the largest and most inclusive: It includes information on more than 5,000 patients from many ethnic groups. Members of the consortium are currently working to design a clinical trial to confirm the results of this prospective study.

"This research study has made an important advance toward personalizing medicine - it uses data from countries around the world to develop a gene-based strategy for warfarin dosing that could benefit a wide range of patients," said Jeremy Berg, PhD, director of the National Institute of General Medical Sciences. "This is a wonderful example of international cooperation and the results are especially valuable for the United States, since our population is so genetically diverse."

Altman, Klein, Stanford medical student Hersh Sagreiya and members of the consortium collected a variety of demographic, clinical and genetic information from patients at risk for stroke, heart disease or other blood clotting problems for whom the ideal doses of warfarin had already been established by trial and error. In particular, they determined which version of each of two genes was carried by each patient. One, called CYP2C9, affects how the liver activates and excretes warfarin. The other, called VKORC1, activates vitamin K, which is essential for blood clotting.

They combined this information with the pre-determined ideal warfarin dose in about 4,000 patients and used it to develop a computerized dose-prediction algorithm. They then tested the new algorithm with and without the genetic data on the remaining 1,000 or so patients.

"We found that even just including demographic and clinical information, such as the patient's height, weight, ethnic background and other medications, yielded initial doses that were closer to the final, ideal dose, than the standard, ‘fixed-dose' regimen," said Altman. "But including the genetic data enabled us to be much more accurate."

Russ Altman, MD, PhD, and Teri Klein discuss their findings and the implications for patients who are treated with warfarin worldwide. Video length: 3 min

Specifically, this pharmacogenetic algorithm yielded predictions that were, on average, within about 8.5 mg/week of the patient's ideal dose; using just the demographic and clinical data predicted doses that were within about 10 mg of the ideal weekly dose. In contrast, starting every patient on 35 mg/week as part of the more standard fixed-dose approach gave an average error of about 13 mg. The algorithm and the data used to derive it will be publicly available on the Stanford-managed Pharmacogenetics and Pharmacogenomics Knowledge Base after the study is published. (Altman is the principal investigator and Klein is the director of the PharmGKB database, which is a curated, international repository for data and knowledge intended to aid researchers in understanding how genetic variation among individuals contributes to differences in reactions to drugs.)

"Right now a physician has no way of knowing, when you walk in to his or her office, whether you will require an unusually high or unusually low dose of warfarin," said Altman. "This algorithm will flag the 10 to 30 percent of people in whom determining the ideal dose will be difficult by standard dosing methods." The researchers are continuing to fine-tune the algorithm in a similar study of about 100 Bay Area patients of the Stanford Anticoagulation Clinic.

Incorporating this kind of genetic information in a standard office visit will be a challenge for many clinicians, Altman admitted. Although many companies are trying to develop an easy-to-use, near instant genotyping test, the time lag between obtaining a DNA sample and getting a useful result can now vary from hours to days. However, the fact that patients are currently recalled for weekly blood tests might make a combination approach workable, said Altman. A physician could start a patient on a standard dose of warfarin and take a DNA sample during the first visit and have the results necessary for fine-tuning by the time the patient returns the following week. And once the genetic information is obtained for a patient, it may also be possible to guide the use of many other drugs.

"Our whole goal is not to have warfarin be the only drug for which this approach is useful," said Altman. "CYP2C9 variants affect the metabolism of about 20 or 30 other commonly used drugs." He and Klein are also working with another consortium, the International Tamoxifen Pharmacogenomics Consortium, to determine the effect of a related protein, CYP2D6, on outcomes for women taking the breast cancer drug tamoxifen.

"It's really an amazing story," said Altman. "Even though we described what is probably the best ever clinical algorithm for warfarin dosing, this paper shows conclusively that including a patient's genetic information yields a far superior prediction. It's a vast improvement over the guessing game physicians play now."

The study was supported by grants from the National Institutes of Health Pharmacogenetics Research Network, the National Institute of General Medical Sciences, the National Heart, Lung and Blood Institute, the National Institute of Neurological Disorders and Stroke, the National Center for Research Resources, the National Research Program for Genomic Medicine and numerous other international granting agencies.

Klein has received consulting fees from Affymetrix; Altman has received consulting fees from and holds equity in 23andMe Inc. Altman is also a member of the Stanford Cancer Center.

About Stanford Medicine

Stanford Medicine is an integrated academic health system comprising the Stanford School of Medicine and adult and pediatric health care delivery systems. Together, they harness the full potential of biomedicine through collaborative research, education and clinical care for patients. For more information, please visit

2023 ISSUE 3

Exploring ways AI is applied to health care