CLN Article
Supported by
As pharmacogenomics (PGx) propels the personalized medicine movement, the ongoing conversation centers around best practices for using the information it produces to enhance patient care.
One of the key players in these conversations is Kelly E. Caudle, PharmD, PhD, BCPS, FCCP. In addition to being an associate member of the St. Jude Children’s Research Hospital faculty, she is also director of the Clinical Pharmacogenetics Implementation Consortium (CPIC), an international group that creates guidelines for pharmacogenetic tests for patient care.
The group aims to deal with the difficulty in translating genetic laboratory test results into actionable prescribing decisions for affected drugs by creating, curating, and posting freely available, evidence-based guidelines.
At CPIC, Caudle coordinates writing all the clinical guidelines. Not only has she identified gaps in testing information, but she also is working to tackle barriers to implementation of this care, from how to best explain the results to patients, to standardizing information for the use of electronic health records, to tackling the persistent issue of getting health insurance companies to pay for testing.
It begins with the patient. Our patients overwhelmingly opt in to genetic testing: 96% of them say ‘yes.’
We have performed pharmacogenomic testing for over 7,400 patients to date. Most of the 4% who say ‘no’ may not be getting major treatments here, but if they are, we treat them as if we didn’t know what their genetic status was.
However, we always test before we prescribe a medication with high risk of toxicity like thiopurines in patients with acute lymphoblastic leukemia. Out of the 4% of patient families declining pharmacogenomic testing, the most common reason for declination is due to patients lack of interest in being genotyped.
Overall, we’ve implemented 16 genes impacting 75 drugs into our electronic health record (EHR) based on the CPIC guidelines.
If a patient has what we call a high-risk result, that means they have some type of variation that may cause them to have increased risk of toxicity to a particular drug or it may have decreased efficacy. When a prescriber orders a medication in the EHR, and the patient has a high-risk result, the prescriber will get a post-test alert. It’s an effective way to warn the provider at the point of care that this patient may have issues with that particular drug, and we also provide the recommendations from the CPIC guideline as part of that alert.
To start, our nurses get consent for testing, and explain what that testing means in patient-friendly terms. We don’t use big medical words and instead make it so that a layperson can understand.
Generally, we tell them that some people have genetic differences that cause them to respond differently to certain drugs and talk them through that process.
After the results come in, those patients can see one of our PGX pharmacists, who can give a detailed interpretation of these results. We tell our patients up front that we will provide additional support for them if they have questions about the testing.
We also provide these results back to patients because the results do not just affect them here at St. Jude if they see another provider. It’s not like a temporal glucose level. These PGx results will affect patients for the rest of their lives.
There are caveats to that, though. Maybe there will be additional variants that we’ve discovered since the time they had that test, so they wouldn’t have been tested for that particular variant.
Acute lymphoblastic leukemia (ALL) is certainly a notable example. We test for multiple variants. With ALL, patients receive a lot of supportive medications for nausea, vomiting and pain, for example.
Many patients also receive an antifungal at some point, and that’s also affected by genetic variation, so we test for that, too.
PGx also is beneficial in other types of cancer. For example, the DPYD gene is important for the metabolism of fluoropyrimidines, medications used to treat colorectal, breast and other cancers. DPD deficiency can significantly increase the risk of severe toxicity in patients treated with 5-fluorouracil or its prodrug, capecitabine.
Consequently, testing for DPYD has become a widespread practice in colon cancer treatment, where 5-fluorouracil is the backbone of many chemotherapy regimens. There have even been lawsuits over not testing for DPYD because patients died in response to those cancer drugs when their risk for taking them would have been spotted by a PGx test.
We always need more research, more literature, and more evidence.
We know a lot about PGx in some populations, like in Whites, African Americans, and Asians, but there are lot of populations out there that have not been studied. We can extrapolate into those other populations, thinking variations in one gene probably affect everybody the same, but we need evidence to be sure that is true.
We want to provide the best care for everyone, and we need more information about other populations to do that, especially as the field grows.
From an implementation standpoint, the big barrier we have is lack of standardization. In the last few years, a lot of my work has focused on standardization in testing. I mentioned earlier that maybe a patient must be retested later in life because the test itself didn’t have enough variants for it to be impactful. Right now, each laboratory’s test is different when it comes to what variants are included and how that information is imported into electronic systems.
This lack of standardization can pose significant problems, especially when integrating this data into the electronic health record (EHR) system. It's crucial that everyone, including our computers, speaks the same “language” to ensure accurate communication and analysis.
There is a big push for PGx testing in the mental health field. At CPIC, we already have guidelines for antidepressants, and we’re working on one for antipsychotics.
We also see PGx growing in cardiovascular medicine. Genes that encode enzymes implicated in the metabolism of warfarin, clopidogrel, and beta blockers are being tested. In addition, proton pump inhibitors used to treat acid reflux are also affected by variations in genes.
There are also quite a few examples in HIV treatment where PGx can be useful. It really does span almost every discipline in some way, especially since a lot of these drugs are used for supportive care.
The data for pediatrics is quite lacking compared with that for adults. It’s a huge gap. But there is some guidance, even if it’s drawn from adults. For drug metabolizing enzymes, we would expect genetic variations to have the same consequence in children older than two years old, but there is a lot more variation in metabolism of drugs in children than adults, so it is important that we still study this vulnerable population.
Often, in our recommendations we don’t expect it to be different in pediatrics, but we acknowledge there is a gap. The exception is ALL, since most of the studies are done in the pediatric population. Those are probably our best guidelines for pediatrics. There’s great data there.
I think with machine learning and artificial intelligence, the field is going to keep growing in new ways. If you’ve had a genetic test, I think at some point, EHRs will have sophisticated algorithms that pulls all your information together to make it more informative—from renal function, liver function, smoking status, to drug interactions -- anything that can affect a dose of a drug or selection of a drug.
One barrier is reimbursement of the genetic test, as some insurance companies are not covering PGx testing. Until we get through some of those issues, PGx testing will not be as widespread as it could be. In 10 years? Maybe. In 20 years? I would love to see that.
Hopefully, we’ll see not only broad access to PGx testing but prescribers who will have technology that can help them think through what to prescribe for that specific patient – that would be great.