The Rise of Precision Prescribing

When University of Florida (UF) Health started recommending genotyping to guide clopidogrel prescribing about ten years ago it was a “life or death issue,” explains Larisa Cavallari, professor of Pharmacotherapy and Translational Research at UF. The drug is administered after angioplasty to prevent stroke. More than half a million angioplasties are performed in the US alone each year.

So, it was strong evidence pointing to the high risk to patients that led to the move. “UF was part of a multicenter trial that found that genotyping helped reduce risk of stroke significantly,” she says. (Cavallari et al.  JACC, January 2018.)

Clopidogrel is the most commonly prescribed P2Y12 inhibitor, or antiplatelet drug, but it is associated with an increased risk of major adverse cardiovascular events in patients with loss-of-function CYP2C19 alleles. This genotype does not impact clinical response to alternatives, prasugrel or ticagrelor. Routine implementation of CYP2C19 genotyping to guide antiplatelet therapy can thus save lives.

Thus, with genotyping for clopidogrel prescribing, UF believed it had to step up to save a significant number of patients’ lives. But today UF’s pharmacogenomics (PGx) efforts are shaped by more than that kind of striking data. “Now our new initiatives are often physician-driven,” Cavallari explains.

For example, “Doctors started coming to us and saying they had patients whose pain they could not control, they had heard that genotyping might help,” she says. Now CYP2D6 testing to guide chronic pain prescribing is available. It has led to improved pain management as well as a reduction in opioid consumption. (Smith et al. Genetics in Medicine. 2019).

“If the physicians want certain testing we can work with them to move it forward,” she says.

Of course, there are other hurdles than just knowing the gene-drug interactions. “You must take into consideration drug-drug interactions, comorbidities, and of course, insurance,” says Jennifer J. Hockings, Pharmacogenomicist Clinical Specialist in Cleveland Clinic’s Genomic Medicine Institute. “If I prescribe someone a drug that is a better fit for them, it’s not going to help them if they can’t afford it.”

The ideal, Hockings and others say, will be to test every patient for a wide range of variants and put that information in the electronic health record. As evidence is built up about whether a gene-drug interaction is actionable, that would be added automatically. In a recent review, Hockings and colleagues noted that more than 90% of patients are thought to carry at least 1 genetic variant that should prompt a change in dosing or medication. (Hockings et al. Cleveland Clinic Journal of Medicine, 2020).

In the meantime, the challenge is just to do the testing that is already advised.

Making Genetic Testing Commonplace

UF is now genotyping for multiple drugs, including clopidogrel, thiopurines, opioids, SSRIs, and proton pump inhibitors. But it’s a complex process. A multi-disciplinary team of experts in diseases, genetics, pathology and clinical bioinformatics work to establish the need for the testing and then to make it a seamless experience for physicians through the electronic health record.

“We flag or send an alert about significant genotyping results,” Cavallari says, “But we don’t notify them about normal results. One of the most important things is not to trigger alert fatigue.”

 At Mayo Clinic, meanwhile, the system has implemented a pharmacogenomics certification program. The curriculum for this one-of-a-kind certificate course guides learners through foundational concepts and advanced decision making with lectures, expert panels, case-based presentations, and patient role-play. Mayo experts discuss the necessary elements for considering pharmacogenomics implementation.

‘I would argue that pharmacogenomics is moving too slowly, says Christopher B. Grilli, Outpatient Pharmacy Manager at Mayo Clinic. “It must be in the curriculum.” But adding anything to a doctor or medical student’s already crammed schedule is challenging.

At Mayo, he explains “We have been having this conversation or years.” After the human genome was sequenced there was an assumption that a lot of this would fall into place. “We wanted an easy button.” Instead they have had to grapple with how to teach a whole range of health care professionals, including pharmacists, nurses, doctors, and dentists about the growing number of gene-interactions with drugs.

“We have panels that screen for 80-90 variants,’ Grilli says “But the research continues and we are sequencing more and more patients.’

Improved matching of medications to patients is the “lowest hanging fruit,” according to Burns C. Blaxall, Director of Precision Medicine at The Christ Hospital Health Network. Non-optimized medication is now the fourth leading cause of death, he points out.

Last year Christ Network partnered with genetic testing company Invitae to provide more routine testing. “We are convinced that this will improve outcomes by 20-40%. Reducing emergency room visits and hospitalizations significantly,” says Burns.

But they are also sensitive to the issue of alert fatigue. “Getting a physician to pay attention to an alert is as tough as getting me to lose 10 pounds,” Blaxall says.  As a result the system is using clinical pharmacists who have done residencies in pharmacogenomics. These pharmacists are embedded in primary care physician offices where they help both implement pharmacogenomic guidelines and inform doctors about new genotyping guidelines.

As Blaxall points out, there has been a stark increase in the number of medications recommended for prior genetic testing by CPIC (Clinical Pharmacogenetics Implementation Consortium). “Studies have shown that one in five new prescriptions should have been guided by genomic testing,” he says.

What’s more, patients are taking more drugs at once, causing further challenges for prescribing.  In a recent study, Invitae and United Healthgroup researchers looked at whether unmanaged pharmacogenomic and drug interaction risk could lengthen hospitalization and might influence severe health outcomes in COVID-19 patients. (Ashcraft et al. J Pers Med, 2021.)

The team reviewed medical and pharmacy claims from 6025 Medicare Advantage members hospitalized with COVID-19. Patients with a moderate or high pharmacogenetic interaction probability (PIP), which indicates the likelihood that testing would identify one or more clinically actionable gene–drug or gene–drug–drug interactions, were hospitalized for 9% and 16% longer, respectively, compared to those with low PIP.

High PIP was significantly associated with longer length of stay compared to low PIP in patients with hypertension, hyperlipidemia, diabetes, or chronic obstructive pulmonary disease (COPD). A greater drug–drug interaction risk was associated with 10% longer LOS among patients with two or three chronic conditions.

“70% of visits to doctors involve drugs,” says Kristine Ashcraft, medical affairs director for Pharmacogenomics at Invitae and lead author on that study. “Many Medicare patients take five or more drugs at one time,” she points out.

Cancer and Depression

Two areas where gene-based prescribing has become prominent are oncology and depression.

The rise of targeted therapy has particularly reshaped oncology, “But still only a fraction of patients get targeted treatment,’ says David Stenehjem, Associate Department Head, Department of Pharmacy Practice and Pharmaceutical Sciences University of Minnesota. “There is an urgent need for new biomarkers.”

For example, the management and prognosis of BRAF-mutant metastatic melanoma has been a key area of change, but also of new opportunity. BRAF/MEK inhibitors, when administered prior to or concurrently with immune checkpoint inhibitors, appear to alter the tumor microenvironment and improve sensitivity to immunotherapy.

“Part of the challenge now is how to rationally combine therapies,” Stenehjem says. Adding a therapy can add additional toxicity. And with immunotherapy there is the additional question of whether the patient will benefit at all. But the upside is so high there is tremendous incentive for it.

“What’s hot right now is moving toward precision-based screening using circulating DNA,’ says Stenehjem. “Using that approach, we could both detect cancer earlier and individualize treatment from the start.”

In depression there is a similar problem. The majority of patients do not respond to the first anti-depressant they are prescribed. But now companies like Myriad Genetics have stepped in with tests.

Myriad’s GeneSight test evaluates how variations in multiple genes may influence outcomes with certain FDA-approved medications commonly prescribed for depression, anxiety, ADHD, and other psychiatric conditions. All known CYP450 and non-CYP450 metabolic pathways of each medication and medication metabolite are weighted for their relative importance.

The company’s research suggests this combinatorial approach is better than single-gene testing at predicting patient outcomes and medication blood levels.

Their GUIDED (Genomics Used to Improve Depression Decisions) study was the largest prospective study to assess the benefit of pharmacogenomics-guided treatment for depression using the GeneSight Psychotropic test versus an active therapy control arm. (Greden et al. Journal of Psychiatric Research. 2019). This was an 8-week, blinded, multi-center, randomized controlled trial of 1,167 subjects with major depressive disorder from 20 academic sites and 40 community sites.

The researchers found statistically significant increases in response and remission rates in the GeneSight arm versus the treatment as usual arm at week 8. Symptoms, response, and remission continued to improve for up to 6 months when patients on a medication with significant gene-drug interactions were switched to one  with no or moderate gene-drug interactions compared to those who remained on medications with significant gene-drug interactions.

Pax Neuroscience, meanwhile, is aiming to go one step further. They aim to develop a test that will confirm a diagnosis of depression and indicate a therapeutic response with a single biomarker. The research builds off of previous humans and animal studies that have shown depression is marked by decreased adenylyl cyclase — a small molecule made in esponse to neurotransmitters, such as serotonin and epinephrine.(Targum et al. Molecular Psychiatry, 2021.)

“This is a reliable biomarker because it depends on platelets, which turnover within a week,” explains Mark Rasenick, President, Chief Scientific Officer, and Founder of Pax. “It could be adapted to high-throughput screening and we could test thousands of samples a day.”

This progress in testing and prescribing for depression could be a major win for the health care system, particularly at this time. The majority of anti-depressant prescriptions are made by primary care doctors, who are managing the COVID-19 crisis at the front lines. Meanwhile, mental health problems such as depression are being exacerbated by the crisis.

The Diversity Challenge

 Besides costs and extra steps, gene tests add another layer of complexity – the hesitance of people to take them. Making patients at ease with genetic testing is thus part of the challenge. For example, doctors seem to believe Black women would be less likely to comply with recommendations for genetic counseling and testing, according to new findings from Washington University School of Medicine in St. Louis published in the Journal of Clinical Oncology.

“Oncologists have very different views of what the barriers to genetic testing are for white and black patients,” says Laura Jean Bierute, senior author of the study and a professor of psychiatry. “Their beliefs may be driving the testing rate.”

This study challenges the notion that black patients are overall just more reticent about genetic testing. “We should not lay all the responsibility for these disparities on the patients,” says Bierute. “We need to take part of the blame.”

The researchers surveyed 277 cancer doctors around the country to learn why referrals are made so much less frequently for Black women. Of the doctors surveyed, 67% were white, less than 4% were Black, almost 59% were female and almost 62% practiced at academic medical centers.

Although fewer than 2% of doctors surveyed said they were less likely to refer a Black patient than a white patient, other research has found that Black patients are being referred for genetic counseling and testing less than 60% of the time that such testing is recommended by National Comprehensive Cancer Network guidelines. That compares with a referral rate of 93% for white patients.

The team asked doctors whether they believed Black patients were more likely than White patients to refuse genetic counseling and testing. Almost 26% said yes. Another 46% of respondents cited cost as a barrier for Black patients and a potential reason not to refer. Almost 59% said that their Black patients were less likely to trust their doctors’ diagnoses and referrals than white patients were.

The survey indicated that 14% of physicians felt their patients, in general — regardless of race — probably would not follow through with genetic testing and counseling recommendations, according to a press release. More than twice as many, 31%, thought their Black patients would be less likely than white patients to comply with their recommendations for genetic counseling and testing.

It’s been established that while they do share breast cancer mutations in common, black women can carry certain mutations that are unique to them. These may, therefore, be usual and cause more consternation than reassurance. Data also suggest Black patients are more comfortable working with providers of the same race, but only about 3% of U.S. oncologists are Black. “Most of these studies have been done on people of European ancestry,” says Melissa B. Davis, of the Weill Cornell Medicine Englander Institute for Precision Medicine. But it’s been established that Blacks may have different variants, and those may not be well recognized. “What is the point of my ordering a test if it will not change my action?”

Davis recently published about two gene variants more commonly found in African American woman that may explain why they are more susceptible to triple negative breast cancer than white women of European ancestry. The team used data from their African-enriched Center for the Study of Breast Cancer Subtypes (ICSBS) cohort. (Martini et al. Scientific Reports. 2021).

This ethnic disparity harkens back to the often-cited example of warfarin, the anticoagulant that was so long held up as an example of why pharmacogenomics was really not that useful. Warfarin is a mainstay antithrombotic, but the optimal dose of this medication varies significantly and at the wrong dose it can cause excessive bleeding, cerebrovascular clotting or stroke. It has been long known that response to this drug can be affected by variants in the CYP2C9 (cytochrome P450 2C9) and VKORC1 (vitamin K epoxide reductase complex, subunit 1). But the trials of genotyping for these variants were overall, inconclusive.

UF’s Cavallari has made warfarin a particular area of study. “It is unfortunate that warfarin was one of the first examples of pharmacogenomics,” she explains. It turns out that a flaw of one of the key studies used to discredit warfarin genotyping was that they did not include variants that occur mainly in Blacks.” (Cavallari, LH, Future Cardiology 2017).

“Don’t use genotyping if you don’t test for the right variants,” Cavallari says. “That’s the correct lesson from the warfarin example.”

Malorye Branca is a freelance journalist who has been covering drug discovery, genomics, and healthcare for 30 years. She has degrees in journalism and biology and has written over 300 news articles, features, and analytical reports. She is the co-author of the 2017 book Moneyball Medicine.