CEO SUMMARY: Medicare’s decision to cease covering many pharmacogenomic tests puts as many as 19 million Americans who have genetic variations affecting their response to medications at risk. These medications are commonly prescribed for patients with cardiovascular disease, pain, depression, anxiety, and cancer. Meanwhile, medical centers such as Mayo Clinic are conducting clinical studies to collect evidence that appropriate use of pharmacogenomic tests can improve patient outcomes while also reducing the cost of care.
WHAT THE FEDERAL GOVERNMENT giveth with one hand, it will often taketh away with the other hand. It might be argued that this is true of federal support of pharmacogenomic testing-particularly for those tests clinical laboratories use to identify how patients metabolize many types of prescription drugs.
The federal government hand that giveth is the $215 million initiative announced by President Obama in January to foster the development of precision medicine. Government officials said that the National Institutes of Health and other departments would use the funds to generate the scientific evidence to move the concept of personalized medicine into clinical practice.
The federal government hand that taketh away on this matter is the Medicare program. Since last fall, the nation’s Medicare Administrative Contractors (MACs) have been discontinuing payment for pharmacogenomic testing that identifies how patients metabolize and respond to prescription drugs. As of June 22, all MACs had stopped paying for these tests. The last one, Noridian, ended payment for these tests on June 22.
“As many as 19 million Americans could be affected by this decision,” stated Kristine Ashcraft, Chief Operating Officer of Genelex, a pharmacogenomic testing company in Seattle, Washington. “That’s because about 75% of Americans have genetic variations affecting their response to medications. These medications are commonly prescribed drugs for patients with cardiovascular disease, pain, depression, anxiety, and cancer. When you consider that 4% of Medicare spending is for hospitalizations caused by adverse drug reactions, taking away coverage for tools that can help combat this problem is short-sighted.”
In its Local Coverage Determination (LCD) effective June 22, Noridian said it would end genetic testing to assess patients taking some medications and that for other medications, it would await definitive utility for such testing. In the LCD, Noridian said it ended payment for all genetic testing associated with all medications related to CYP2C9 (CPT 81227) and VKORC1 (CPT 81355). For genetic testing for medications related to CYP2C19 (CPT 81225) and CYP2D6 (CPT 81226), Noridian ended payment until definitive clinical utility is established. Testing for response to medications related to CYP2C19 and CYP2D6 would be limited for patients with certain indications, Noridian said.
“This decision is designed to save money at the expense of patient care,” said John Logan Black, III, M.D., Co-Director, Personalized Genomics Laboratory and Vice Chair for Business Development in the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester, Minnesota. “It comes down to dollars and cents and not really what patients need,” he explained. “Medicare officials are trying to control costs by saying they haven’t seen enough benefit from these tests.
“However, it has long been considered that pharmacogenomic testing is the low-hanging fruit from the Human Genome Project, because these tests give us a lot of personalized information about patients,” he commented. “But now critics say it’s too hard to implement or insurers say they’re worried about the costs to implement this testing.”
Black pointed out that, even as Medicare ends or limits such coverage, the federal Food and Drug Administration recommends pharmacogenomic tests for medications. “Not only is such genetic testing recommended, but, in many cases, FDA clearance requires pharmaceutical companies to print those recommendations for genetic testing on the product information forms patients receive when prescribed these medications,” he explained.
“The FDA’s website shows the Table of Pharmacogenomic Biomarkers in Drug Labeling,” he continued. “This is a list of medications that pharmacogenomic biomarkers affect. These are biomarkers- meaning pharmacogenomic targets-that the FDA requires in drug labeling.
“On the list are a number of cytochrome P450s that require disclosure on drug labels as either cautionary, or, in some cases, the drug label contains information that says, basically, ‘If you’re going to use a specific dose, you (meaning the treating physician) should test this cytochrome P450.’ In some cases it says you should just test cytochrome P450 regardless of dosing,” he said.
Providers Caught in Middle
“So now what happens when Medicare payment doesn’t follow those recommendations?” Black asked. “It puts healthcare providers in a difficult position because-on one hand the FDA says, ‘You should test this patient’-and on the other hand, Medicare says, ‘We’re not going to pay for it.’ The patient gets caught squarely in the middle and may end up paying for that genetic test out-of-pocket. Or, if the patient decides not to be tested, then that patient may be at increased risk.
“The FDA is an important and reliable source of information on these biomarkers and how they interact with some of the most common medications,” continued Black. “But in addition to what the FDA says, the recommendations of the Clinical Pharmacogenetics Implementation Consortium should also be considered. CPIC publishes articles used to set practice guidance risk. This information is based on published literature.
“After looking over the literature, this committee evaluates the research on these various medications and biomarker studies and publishes practice guidance on the use of pharmacogenomic tests,” he stated.
“For issues in which clinicians need practice guidance, CPIC will put together a group of experts who write and publish a paper on the issue. For many of us, that paper then becomes the practice guidance for the use of pharmacogenomic tests in clinical practice.
“So once again, clinicians have a conundrum because these guidance documents are published, highly-regarded expert opinions,” emphasized Black. “Many of us in the field consider them to be the guidance documents for pharmacogenomics.
“So now this best-practice guidance is published and available to all,” said Black. “Therefore, I can envision a physician potentially being involved in litigation where the patient alleges that the physician didn’t follow what is considered ‘practice guidance’ in this area and the patient ended up with a horrible side effect. I’m not aware of any case like that, but it’s possible.”
Up to this point, Black was discussing the most common type of pharmacogenomic testing, which is reactive testing. “A common example of reactive testing is when a patient with heart disease gets a stent and the cardiologist prescribes clopidogrel, which is an anti-platelet agent, and the gene involved is CYP2C19,” he said.
“The FDA has issued a boxed warning for the brand name of this drug, Plavix, and it has issued precautions on this medication regarding dosage and administration,” stated Black. “A boxed warning is FDA’s most serious warning. It says the drug has diminished effectiveness in those individuals who are CYP2C19 poor metabolizers.” (See boxed warning in the sidebar on page 12.)
“Normally, a patient would get this genetic test after starting therapy and the indication became apparent to the physician,” Black explained. “So reactive testing is one way to proceed.
“Now, what some medical centers, including Mayo Clinic, are moving toward is proactive testing,” noted Black. “When a patient falls into a demographic group that has increased risk factors indicating the possibility of developing heart disease, as an example, the physician may want to do the CYP2C19 or other testing long before the patient actually needs medication.
“Then, in the future, if and when that patient needs a stent, the physician knows the most appropriate drug to use,” he said, “That is because an alert fires in the medical record automatically, telling the clinician immediately which medication to prescribe. “This prevents a patient from getting a
medication that otherwise may be as effective as water,” emphasized Black.
“This method of testing puts these patients on the correct treatment regimen the first time to protect them from premature stent closure and possible cardiac death. That’s one example of the clinical value of these tests.
Mayo Physician Says: ‘This Is Not Some Cool Idea. Pharmacogenomic Tests Improve Patient Care’
HERE’S A STORY ABOUT A REAL PATIENT WHO would have benefited from proactive pharmacogenetic testing, said John Logan Black, III, M.D., Co-Director of the Personalized Genomics Laboratory at Mayo Clinic. “Unfortunately, this case is not unusual,” he stated.
“I heard from a clinician that a patient came in from a nursing home after having a pulmonary embolism,” recalled Black. “The doctor prescribed warfarin and sent the patient back to the nursing home. When the patient returned a week later, his INR was off the chart, and it wouldn’t come down. And the patient was coughing up blood.
“To reverse the effects of warfarin, the physician prescribed vitamin K, and the INR came down a bit, but wouldn’t stay down,” he said. “Now the clinician was concerned that the patient got rat poisoning in the nursing home, so he considered filing a report of a case of a patient getting rat poison. But before he did, he ordered the CYP2C9/VKORC1 test for warfarin sensitivity.
“The result showed that the patient had the worst genotype possible for taking warfarin. Thus, he’s going to be at high risk for bleeding,” noted Black. “This patient will have a very high INR on the usual doses of warfarin. It may be acceptable to give him maybe one-fifth of a normal dose. But the physician would need to be extremely careful with any dosage of warfarin. It might be better to prescribe a different drug.
“The point of this story is that pharmacogenomic testing produces real results that make a significant difference for people,” concluded Black. “This is not some kind of genetic testing that we think might be a good idea to try. It’s not something someone thought would be cool to do because we have the ability to assess genetic information. Every day, I see how these pharmacogenomic tests improve patient care.”
Higher Risk Of Bleeding
“A physician could do proactive testing on patients who need warfarin,” he continued. “Should I see a genotype indicating high risk, then I have to call the treating clinician to warn that this patient is at a high risk of having a very high INR-and a higher risk of bleeding-if that patient gets the usual doses of warfarin.
“For these cases, we typically don’t do proactive testing and so patients are put on warfarin,” he said. “Then, when these patients have problems, the physician must determine what happened. With Medicare no longer reimbursing for these pharmacogenomic tests, the question becomes,
‘Who will pay for this genetic testing for a patient in a nursing home?’
“Some patients who are nursing-home bound will be able to pay and some won’t,” observed Black. “Basically, these Medicare patients will be denied access to personalized medicine at the same time that there is a presidential initiative to promote precisely these genetic tests.
“Here at Mayo, we recently conducted a study involving about 1,000 patients,” stated Black. “Each study participant was genotyped for five genes (CYP2C9, CYP2C19, CYP2D6, SLCO1B1, and VKORC1).
“All of these genetic test results were put into Mayo Clinic’s electronic health record system to help clinicians who would manage these patients. And, by the way, this testing was free to the patients,” he added. “We found that only 20.5% of patients who had our CYP2D6 testing did not have a variant! The remainder of those 1,000 patients-meaning 79.5%-had an issue with CYP2D6.
“Further, just 40.1% of these patients had no CYP2C19 variants, leaving 59.9% with an issue,” said Black. “For CYP2C9, 63.7% had no variants and the rest had issues. This tells us these are very common variants.
“When all of the variants found in our samples are added up, only 1% of those 1,000 patients had no variants for all of the genes,” he continued. “The remaining 99% had some genetic variants, which might impact their care depending upon which medications they will need in the future.
“Now take the results of this study and view the findings as a health insurer would view them,” suggested Black. “An insurer may state that, ‘If only 1% of patients were normal across the board, this would indicate that everyone needs this testing before prescribing these drugs. But our health plan cannot afford that.’ This is what is happening when Medicare says it won’t pay for these genetic tests. It is reacting to the costs.
“When you pick a population at risk, what is the collective value of this testing?” asked Black. “To answer that question, you have to ignore the concerns of the individual patient whose stent will close because he/she got the wrong drug. Or a patient will have a major heart attack.
“So, insurers are asking, ‘What was the cost of treating that patient versus the cost of pharmacogenomic testing?'” noted Black. “Apparently, what insurers see-at least at this moment-is that this genetic testing doesn’t pencil out from a financial perspective.
FDA Recognizes Need for Pharmacogenomics
ON ITS WEBSITE, THE FOOD AND DRUG ADMINISTRATION publishes a table of pharmacogenomic biomarkers used in drug labeling. The table lists FDA-approved drugs with pharmacogenomic information in their labels.
On the list are 171 prescription medications. The labeling for some, but not all, of the medications includes specific actions that clinicians should take based on the biomarker information.
“Pharmacogenomics can play an important role in identifying responders and nonresponders to medications, avoiding adverse events, and optimizing drug dose,” the FDA site says. When the FDA requires a warning on a drug’s product label, the warning may contain information on drug exposure and clinical response variability, the risks for adverse events, genotype-specific dosing, mechanisms of drug action, and polymorphic drug target and disposition genes.
The biomarkers the FDA lists include germline or somatic gene variants, functional deficiencies, expression changes, and chromosomal abnormalities. The FDA website also lists some protein biomarkers used to select patients for treatment.
“The FDA’s biomarker list demonstrates how many indications for pharmacogenomic testing are currently identified and considered clinically relevant,” noted John Logan Black, III, M.D., Co-Director, Personalized Genomics Laboratory at Mayo Clinic. “Some of those on the list may be a bit soft and that reflects caution on the part of the FDA.
“But for other biomarkers, there is a strong indication that genetic testing is needed,” commented Black. “For example, clinicians need to consider checking CYP2C19 if a patient is prescribed the antidepressant citalopram at doses greater than 20mg due to increased risk of a cardiac condition known as long QT syndrome. The physician would not know that fact if this genetic test were not performed.”
Collecting Data for Coverage
“Now that we have the results of the five gene-panel analysis on 1,000 patients, we’re considering expanding that work to 10,000 Mayo Clinic patients,” he said. “We would follow them and try to produce numbers to show insurers that this testing could be reasonable to cover.
“We feel it’s important to put the patient at the center of the entire medical discussion and back that position with big dollars,” noted Black. “I don’t know exactly how much was spent to run the five genepanel test for the study. We also have a nine-gene panel that we offer internally here, and when we run that test, those results go into the patients’ medical records so that the results are available for their future care.
“It’s important to note that Mayo Clinic is not the only medical center doing this type of testing,” concluded Black. “Others are doing some proactive testing so that they will have those results for their patients for years to come.”
Tamoxifen, Codeine Have a Useful CYP Gene Test
TAMOXIFEN IS ONE OF THE MOST COMMON MEDICATIONS for which physicians use pharmacogenomic testing of CYP2D6 to assess how well a patient will metabolize the medication.
“The test for CYP2D6 assesses for responsiveness to tamoxifen, an agent used to prevent breast cancer reoccurrence. CYP2D6 turns tamoxifen on,” said John Logan Black, III, MD, Co-Director of the Personalized Genomics Laboratory at Mayo Clinic. “So, it’s very important to know how the patient will respond or if the patient will respond to tamoxifen therapy.
“For women who are prescribed tamoxifen, some will get tamoxifen and have a recurrence of breast cancer because of the way they metabolize the medication,” he added. “These women would have better outcomes if the genetic testing were done and they didn’t have a recurrence of breast cancers.
“One drug for which the CYP2D6 test is even more important is codeine, a very common analgesic,” he stated. “If the patient metabolizes codeine poorly, then it is like taking water. It won’t help the patient. But if the patient is an ultra-rapid metabolizer and takes the typical dose of codeine, the patient may overdose because the patient converts codeine to morphine too rapidly. Some patients will hallucinate, and others may have respiratory depression. Pediatric patients appear to be at particular risk.
“Many medical centers will genotype all pediatric patients upon admission to look for CYP2D6 variants,” noted Black. “Or, most medical centers have decided not to use codeine because of this risk. So, for codeine, the results of this pharmacogenomic testing are very significant.”
Contact Andy Tofilon, Mayo Medical Laboratories at email@example.com or 507-538-5245; Kristine Ashcraft at 800-837-8362 or firstname.lastname@example.org.