Originally published Sept. 10.
Since adding a new immunosuppressive drug, called tacrolimus, to its pharmacogenetics program earlier this year, Vanderbilt University's doctors have used the school's electronic medical records to identify several thousand patients who harbor gene variations that compromise their ability to respond to this treatment and place them at risk for organ transplant rejection.
Marketed under the brand name Prograf, tacrolimus is given to patients after they've received a transplant to reduce the activity of white blood cells trying to rid the body of the new organ. The drug is traditionally dosed based on a patient's weight, but this can yield variable effects, because tacrolimus has a very narrow therapeutic window.
Too little of the drug can cause a patient's body to reject the transplanted organ and too much can cause diabetes or certain types of skin cancer. Within the first year after organ transplantation, 10 percent of patients treated with tacrolimus experience rejection. Approximately 33 percent of patients within the first year of an organ transplant are newly diagnosed with diabetes, which is partly attributable to inaccurate tacrolimus dosing.
After Vanderbilt's institutional review board green-lighted gene markers linked to tacrolimus response based on the published literature, the university in March began reporting genetic data associated with drug response to patients via its EMR system as part of its PREDICT pharmacogenetics testing program. Once reported as part of patients' EMRs, doctors can use this information to inform their prescribing decisions.
Kelly Birdwell, an assistant professor of medicine at Vanderbilt who is conducting research on tacrolimus pharmacogenetics, told PGx Reporter that the PREDICT program now reports CYP3A5 *3 and *1 alleles, as well as less-common markers, such as the *6 allele.
Last month, the university reported that since launching PGx testing for tacrolimus, more than 2,800 Vanderbilt patients have been found to carry such gene variations. Of these patients, 600 have received or are waiting for heart or kidney transplants.
Past studies have found that CYP3A5 polymorphisms are associated with concentrations of tacrolimus in the blood, how quickly patients with this genotype clear the drug, and the dose they need to avoid transplant rejection. In a genome-wide association study published in Pharmacogenetics and Genomics last year, researchers led by Birdwell found in accordance with other studies that those carrying CYP3A5 *1 alleles cleared tacrolimus more rapidly, had lower concentrations of the drug in their blood, and needed higher doses than patients who were homozygous for the CYP3A5 *3 allele.
In the study, researchers genotyped around 400 patients for nearly 1,000 polymorphisms, and replicated previously described associations between the CYP3A5 rs776746 SNP and the tacrolimus dose requirement, considered in terms of the concentration-to-dose ratio. When adjusted for other variables, such as age, sex, and weight, CYP3A5 rs776746 was responsible for nearly 40 percent of the variability among patients taking the required tacrolimus dose.
The CYP3A5 rs776746 SNP encodes the non-functional CYP3A5 *3 allele. Carriers of the CYP3A5 *3 allele tend to build up higher levels of tacrolimus compared to *1 carriers, and therefore require a reduced dose.
In the same study, Birdwell and colleagues also identified nine other polymorphisms in chromosome 7 that were less strongly linked to the tacrolimus required dose, but were in linkage disequilibrium with CYP3A5 rs776746.
In order to conduct this investigation, researchers utilized Vanderbilit's DNA biobank, BioVU, which has to date accrued data on approximately 170,000 consenting patients. The data stored in his repository is de-identified and used to conduct large-scale genomic investigations. Eventually, based on the available literature on a particular gene-drug association, Vanderbilt's IRB decides whether or not to include it as part of the roster of markers that will be reported to patients treated at the university as part of the PREDICT program and in their EMRs.
Since launching PREDICT in 2010, Vanderbilt has genotyped more than 14,000 patients in order to gauge how genetic variations impact their ability to respond to treatments. According to the university, more than 12,000 patients, or 88 percent, have genetic markers that increase their chances of experiencing adverse reactions for the five drugs currently included in the program.
Guidelines to come
In addition to tacrolimus, Vanderbilt's PREDICT program reports PGx markers for four other drugs, including the antiplatelet drug clopidogrel, the anticoagulant warfarin, the cholesterol-lowering drug simvastatin, and thiopurine drugs for cancer and autoimmune disorders.
Of these treatments, the FDA recommends genetic testing in the labels for clopidogrel and warfarin. The "clinical pharmacology" section of the label for the thiopurine drug Imuran (azathioprine) informs doctors and patients that the TPMT gene is involved in drug metabolism and that genetic testing can identify patients with polymorphisms in this gene who are at increased risk of myelotoxicity.
Simvastatin's label warns doctors about the risk of myopathy with the drug, and recommends against prescribing patients the 80 mg dose. The label, however, doesn't recommend testing patients for the SLCO1B1 gene, which is associated with a heightened risk of developing severe, statin-induced myopathy. Studies have shown that statin users who are heterozygous for the SLCO1B1 risk allele are about five times more likely to suffer myopathy, while those who are homozygous for the risk allele are about 18 times more likely to experience a side effect.
For tacrolimus, the FDA does not yet recommend testing for CYP3A5 polymorphisms to avoid risk of transplant rejection and other adverse events. However, the Clinical Pharmacogenetics Implementation Consortium is planning to issue guidelines on how pharmacogenetic information may be used to guide tacrolimus prescribing.
Mary Relling, chair of the pharmaceutical department at St. Jude Children's Research Hospital and head of CPIC, told PGx Reporter that the group has formed a writing committee toward this end. CPIC, jointly formed in 2009 by PharmGKB and the Pharmacogenomics Research Network, issues peer-reviewed guidelines focused on how genetic information should be used to guide treatment decisions, rather than on whether tests should be ordered in the first place.
Tracking outcomes
While guidelines on how and when to utilize genetics data is key to spurring adoption of pharmacogenetic testing, the biggest barrier to more widespread use has been a dearth of data showing that testing improves patients' outcomes and lowers costs.
On the cost front, by using its EMR system, Vanderbilt is aiming to change the incentives for incorporating PGx information into medical care. Because Vanderbilt tests consenting patients on a multiplex assay for 184 common polymorphisms in 34 genes at the outset of their care, this PGx information is already stored in a database, although not reported through the EMR until the IRB OKs a particular gene-drug association. However, this system ensures that when the evidence base supports reporting genetic information for medical decision making, doctors will automatically have this data at their fingertips.
“Let's pretend we live in a world where the [genetic] information is already there in the EMR,” posited Dan Roden, assistant vice chancellor for personalized medicine at Vanderbilt. “What would you do with it? It's very different question than 'would you do the genotyping?'”
The University of Florida has taken a similar approach. Hypothesizing that doctors will be more likely to adopt genetic testing in their practices if the data is readily available at the point of care, within patients' EMRs, the UF Health Personalized Medicine Program launched a strategy a year ago in which all patients visiting its catheterization lab would receive a multi-gene test to determine if they are likely to respond to clopidogrel. The test gauges a range of markers associated with various cardiac conditions, but UF is currently only reporting PGx data for clopidogrel within its EMRs and using the other information for research (PGx Reporter 8/21/2013).
UF and Vanderbilt's programs also enable them to collect longitudinal outcomes data to see if pharmacogenetic testing actually made a difference for the patient's health. Specifically, with regard to tacrolimus, researchers at Vanderbilt are using the BioVU DNA databank and other resources to assess whether genetic testing actually helps lower post-transplant complications like acute rejection, calcineurin inhibitor toxicity, squamous cell skin cancer, and new onset diabetes.
“We are monitoring for the same outcomes in a separate prospective group of patients from whom we also collect DNA,” Birdwell said. “In addition, we will monitor the patients under the PREDICT program.”
With the help of Vanderbilt's EMR system, researchers are able to track a variety of data points. For example, researchers are documenting how many patients in its system harbor a particular genotype and how often doctors are paying attention to the EMR alerts to change medications or dosing. Roden said researchers are collecting this type of information on clopidogrel phamacogenetics for future publication.
Based on data collected so far, Roden noted that when patients don't have loss-of-function CYP2C19 markers, 95 percent end up taking clopidogrel, which is now generic, while the majority of those who are homozygous for the risk allele end up on another drug. “When you know the genetic variants [for patients], you have the benefit of a cheap and effective drug for those who don't have the [loss-of-function] markers,” Roden said.
Tacrolimus is also available in generic form. As such, more accurate dosing with the help of pharmacogenetics information could help avoid adverse events and potentially yield healthcare savings.