Posted by Sappho on November 27th, 2013 filed in Health and Medicine, News and Commentary
“Warfarin (Coumadin) Sensitivity: Increased” Thus reads one of the lines in my 23andme drug response report. And it’s this line that’s one of the bones of contention between the FDA and 23andme.
Most of my and my family’s 23andme health results are not, in fact, all that actionable.
My mother’s at reduced risk for macular degeneration and Alzheimer’s. Yay! But, since Mom is nearly 80, the fact that she’s shows as at reduced risk for these diseases is evidence that 23andme, in this case, got right the fact that Mom sure enough doesn’t yet have macular degeneration or Alzheimer’s than a solid prediction that she’ll be spared in the future. Once you reach 80, all bets are off on whether your genes will continue to protect you from any age related illnesses at all. The one condition for which Mom’s at greatest increased risk is, sure enough, something that already happened to her, so that 23andme, here, doesn’t tell us anything that family medical history didn’t tell us already.
I’m at increased risk of stomach cancer. But that increase is a jump from a 0.08% chance that I’ll get stomach cancer by the time I’m 80 to a whole 0.07% chance. I think it’s time to start making out my will.
I’m also at somewhat increased risk of atrial fibrillation and at somewhat reduced risk of coronary heart disease. Both conditions are so common that my reaction to the news of my not all that altered risk should be:
Get regular physical activity
Eat a heart-healthy diet, low in saturated fats, trans fats, and cholesterol
Manage high blood pressure
Avoid excessive amounts of alcohol and caffeine
Maintain a healthy weight
In other words, exactly the same things I already should be doing. And the same things that I should have been doing if the increased risk had been for coronary heart disease and the reduced risk for atrial fibrillation, since the two conditions have basically the same lifestyle recommendations for prevention.
Even my health insurance company needn’t care much about my new risk assessment, since the increased and decreased risks cancel each other out, and in any case are far outweighed by the fact that I’m a cancer survivor. The particular cancer that I survived isn’t one of the ones for which 23andme provides a risk assessment, and, in any case, the genetic counselor that I saw after my treatment concluded, after reviewing my family history, that mine was unlikely to be a genetically determined cancer. (No, I didn’t smoke, and no, it wouldn’t have mattered for this particular cancer. Sometimes cancer just happens, and you deal with it as it comes.)
It’s reassuring to find, among my obesity gene results, the one that say that I’m predisposed to respond well to a Mediterranean diet. Great! My best obesity prevention strategy is the one that involves something I already enjoy. Pass the olive oil.
The lab is used by my doctor as an actual diagnostic to decide on real medical treatment.
The 23AM report is used by me to go “wow, I have 3% Mongolian ancestry”.
But that’s before we get to warfarin. If my other health results basically tell me to do what I should have been doing anyway, my drug response results are another story. Most of my 23andme results involve typical response, but several stand out:
- Proton Pump Inhibitor (PPI) Metabolism: Rapid. I carry a variant of a gene called CYP2C19 that makes me likely to be either an ultrarapid or a rapid metabolizer of PPIs, drugs that limit the production of stomach acid. If I take drugs like omeprazole (Prilosec®), lansoprazole (Prevacid®), rabeprazole (AcipHex®) and esomeprazole (Nexium®), particulary if I’m being treated for H. pylori infection, I may need a different dose from most people.
- Warfarin (Coumadin®) Sensitivity: Increased. I have a genotype CYP2C9 *1/*2, VKORC1 -1639/3673 AG, in the CYP2C9 and VKORC1 genes tested by 23andme. This means, according to 23andme, that I have “Increased warfarin sensitivity. May require decreased warfarin dose.” (I could also have come out with either “slightly increased” or “substantially increased” sensitivity, but I fell, apparently, right down the middle, among people with increased warfarin sensitivity.)
- Sulfonylurea Drug Clearance (Type 2 Diabetes Treatment): Reduced. I have “somewhat reduced ability to clear sulfonylurea drugs from the body,” a fact that might be relevant should I ever be diagnosed with Type 2 Diabetes.
- Pseudocholinesterase Deficiency: Increased. 23andme reports on “Mutations in the BCHE gene, which encodes the pseudocholinesterase enzyme, can cause pseudocholinesterase deficiency. 23andMe reports data for three of the most common BCHE mutations associated with pseudocholinesterase deficiency.” I have “One copy of one of the BCHE mutations reported by 23andMe. Increased risk for slightly extended paralysis and apnea after treatment with choline ester drugs.” This puts me at increased risk of pseudocholinesterase deficiency, a condition that can also be caused by “pregnancy, malnutrition, chronic infections, cancer, liver disease, extensive burn injuries and use of certain medications.”
- Phenytoin (Dilantin®) Sensitivity (Epilepsy Drug): Increased. It’s that CYP2CP9 gene variant again. This time it makes it likely that I’ll metabolize Dilantin more slowly.
I’ve taken one of the drugs on the list above, exactly once. It was Prilosec, for about a week. It worked, despite my apparently being genetically predisposed to be a rapid metabolizer.
I’ve also acted on one of the drug response conditions reported. People with pseudocholinesterase deficiency, my 23andme report says, are at increased risk of a heart attack if they take cocaine. I took the opportunity to post on Facebook warning all my nephews and nieces never to take cocaine, since they might be genetically predisposed to heart attacks from it.
More to the point when it comes to medical decision making (since doctors aren’t likely to give me cocaine), I looked up the uses of choline esters and the effects of pseudocholinesterase deficiency, according to the NIH.
Pseudocholinesterase deficiency is a condition that results in increased sensitivity to certain muscle relaxant drugs used during general anesthesia, called choline esters. These fast-acting drugs, such as succinylcholine and mivacurium, are given to relax the muscles used for movement (skeletal muscles), including the muscles involved in breathing. The drugs are often employed for brief surgical procedures or in emergencies when a breathing tube must be inserted quickly. Normally, these drugs are broken down (metabolized) by the body within a few minutes of being administered, at which time the muscles can move again. However, people with pseudocholinesterase deficiency may not be able to move or breathe on their own for a few hours after the drugs are administered. Affected individuals must be supported with a machine to help them breathe (mechanical ventilation) until the drugs are cleared from the body.
People with pseudocholinesterase deficiency may also have increased sensitivity to certain other drugs, including the local anesthetic procaine, and to specific agricultural pesticides. The condition causes no other signs or symptoms and is usually not discovered until an abnormal drug reaction occurs.
I’ve had more occasion to receive general anesthesia since being diagnosed with cancer than in the whole rest of my life. I already had a phobia of general anesthesia, and being treated for cancer is an experience full of scary things. Taking the 23andme test after my cancer diagnosis, I was not going to miss an opportunity to make one of those scary things less scary. So, “I’ve gotten a genetic test result indicating that I may have a pseudocholinesterase deficiency” is now part of my standard instructions to the anesthesiologist any time I receive general anesthesia. “Pseudocholinesterase deficiency” is also included on my MedicAlert necklace (the one I was going to get anyway, even without that 23andme result, because of the cancer diagnosis).
And here is the core of the FDA’s dispute with 23andme. The FDA wants 23andme to meet FDA medical device standards for its risk evaluation, because the FDA considers that risk evaluation tantamount to giving medical advice.
The FDA writes, in its warning letter:
For example, false genotype results for your warfarin drug response test could have significant unreasonable risk of illness, injury, or death to the patient due to thrombosis or bleeding events that occur from treatment with a drug at a dose that does not provide the appropriately calibrated anticoagulant effect. These risks are typically mitigated by International Normalized Ratio (INR) management under a physician’s care. The risk of serious injury or death is known to be high when patients are either non-compliant or not properly dosed; combined with the risk that a direct-to-consumer test result may be used by a patient to self-manage, serious concerns are raised if test results are not adequately understood by patients or if incorrect test results are reported.
Ronald Bailey, at the Libertarian web site Reason, is impatient with the FDA’s argument.
What the test results would actually lead patients to do is to get another test and to talk with their physicians. The FDA also cites the genotype results that indicate the sensitivity of patients to the blood-thinning medication warfarin. Again, such results would be used by patients to talk with their doctors about their treatment regimens should the time come that they need to take the drug. In fact, in 2010 the FDA actually updated its rules to recommend genetic testing to set the proper warfarin dosages for patients.
23andme, having different institutional incentives than Reason magazine, gives a more penitent response. Anne Wojcicki, co-founder of 23andme, writes on the 23andme web site.
… It is absolutely critical that our consumers get high quality genetic data that they can trust. We have worked extensively with our lab partner to make sure that the results we return are accurate. We stand behind the data that we return to customers — but we recognize that the FDA needs to be convinced of the quality of our data as well.
In 2008 we began our dialogue with the FDA. The relationship with the FDA remains critically important to 23andMe.
In July 2012 23andMe submitted its first application for FDA clearance and followed on with another submission at the end of August. We received feedback on those submissions and acknowledge that we are behind schedule with our responses.
This is new territory for both for 23andMe and the FDA….
What I can’t tell, either from the FDA’s account that
As part of our interactions with you, including more than 14 face-to-face and teleconference meetings, hundreds of email exchanges, and dozens of written communications, we provided you with specific feedback on study protocols and clinical and analytical validation requirements, discussed potential classifications and regulatory pathways (including reasonable submission timelines), provided statistical advice, and discussed potential risk mitigation strategies.
or from 23andme’s acknowledgement that “we are behind schedule with our responses,” is exactly what the FDA wants 23andme to prove, that 23andme has not yet proven. Several possible concerns come to mind.
- If 23andme is to tell me that I have genotype CYP2C9 *1/*2, VKORC1 -1639/3673 AG, which may increase my sensitivity to warfarin, then 23andme ought to be able to prove the reliability of their gene sequencing. It seems unlikely to me that this would be the hold up. 23andme has had ample time to supply the FDA with proof of the reliability of their genotyping technology, the Illumina OmniExpress Plus, and genotyping is, by this point, a matter of reasonably well established science.
- The FDA could expect 23andme to validate that the genes on which they’re reporting actually have the health risk effects that they’re reporting. What’s the evidence tying CYP2C9 and VKORC1 mutations to warfarin sensitivity? I’m not sure how this would work, since 23andme reports on many genes, some of them (by 23andme’s own accounting) only tied to conditions by preliminary research, and others (such as BRCA1/BRCA2 variants) already covered by extensive established research.
- The FDA could want reassurance on how 23andme’s customers are reading 23andme’s messaging. My warfarin report from 23andme says “Only a medical professional can determine the right dosage of warfarin for a particular patient. Clinical information such as age, size and other medications the patient is taking can affect a person’s optimal dose. The amount of vitamin K in the diet is also a factor. In addition to the genetic variations reported here, there may be other genetic factors that impact warfarin response. The information contained in this report should not be used to independently establish a warfarin regimen, or abolish or adjust an existing course of treatment.” But do 23andme’s customers, as a group, properly read and understand 23andme’s caveats?
I’m guessing that the FDA and 23andme are sorting out some version of my second item: How much evidence does 23andme need to supply that it’s accurately reporting the risks associated with particular genes? What standards do we apply, in approving genetic tests for health risks? It’s this question that the United Healthcare report last year discussed:
However, of the roughly 1,000 to 1,300 newer and more complex tests, only a minority have demonstrated clinical utility so far. New research models may provide alternatives to traditional clinical trials (such as randomized controlled trials) that include a less expensive mechanism for evaluating genetic and molecular diagnostic tests. Examples of possible models include those that involve rapid iterative cycles, practice-based interventions, observational studies, prospective and retrospective studies, and comparative effectiveness research (CER).
In the case of 23andme’s report on warfarin sensitivity, it seems that the burden of proof ought to be manageable. The FDA has already cleared a genetic lab test for warfarin sensitivity, back in 2007, based on “variants of two genes, CYP2C9 and VKORC1,” that are exactly the two genes on which 23andme reports. In principle, there should be no reason that 23andme can’t meet the same rules that cleared that other lab test. But it’s possible that the larger number of conditions on which 23andme reports makes for more FDA required studies, and more difficulty meeting the requirement.
23andme isn’t the only company facing this issue. Myriad, which has been in the genetic testing business in a more high end, medically prescribed way, only this year applied for FDA approval for its BRACanalysis.
23andme, though, has a different corporate model than Myriad, and so much of the public debate over the FDA’s warning concerns, not the scientific validity of 23andme’s tests, but the question of what happens when you reveal genetic risks to people who may lack the background to put those risks in perspective. A friend passes on to me this article by Steven Novella on the FDA and Personalized Genetic Testing
… This issue aside, meaning even if we had a 100% accurate test, we have to ask the further question – what is the net effect of providing this information? Potential negative consequences of accurate screening information include the unnecessary worry of learning that you have a predisposition or risk of developing a disease in the future. This may seem like a small thing, but it can have a significant impact on quality of life, even to the point of people being crippled by their fear of impending disease.
Further we have to consider what, exactly, will be done with the information. The implication is that drug treatment can be modified, and risk factors for disease can be addressed. But – is genetic testing (with our current knowledge base) the best way to approach such treatment? Will the genetic information cause people to change their behavior in a way that will have a net negative effect on their health?
Such behavioral changes may also include a false sense of security in those with negative screening for certain risk factors. If someone finds that they are in a low-risk group for heart disease, will they then feel free to engage in an unhealthy diet that will still be a risk factor for them?
Screening indicating a high risk may also motivate people to seek out unnecessary health care, which may lead to unnecessary testing that have their own false positive and negative rates….
In contrast, Razib Khan, at Slate, argues
This brings us to the fact that 23andMe is just part of a broader movement toward the democratization of health information. This incident highlights the tension between the paternalistic medical establishment that arose to deal with the dangers of 19th-century quack medicine, and a “techno-populist” element of American society pioneering personal health assessment and decision-making by leveraging new information technologies. Caught between them is the general public, which trusts institutions but is obviously intrigued by the offerings of consumer medicine, as judged by units sold of 23andMe’s kits.
Genotyping whole genome sequencing services are soon going to be as ubiquitous as white bread. The likelihood that the FDA would ban you from reading your raw results seems low. Rather, their concern is when a firm like 23andMe interprets those results. The glaring weakness in an aggressive strategy against interpretative services is that there will always be firms such as 23andMe, and there’s no reason that they need to be based out of the United States….
I am pretty confident of my own ability to understand my drug response results. My mother has a Ph.D. in pharmacology and taught at a medical school, and I can phone her for advice any time I hit something I don’t understand. I’ve shared all my 23andme health reports with her in any case. My sister Carey had Joanna Mountain, 23andme’s research director, as her college roommate. (I attended Stanford with them, and picked Carey and Joanna to supervise me when I went through my ritual of getting drunk to celebrate my twenty-first birthday; I visited both of them when they served together in the Peace Corps in Kenya after college.) The two have stayed in touch over the years, and so, if I were really puzzled over that 23andme warfarin result, I could call Carey, who might say, “I’ll ask Joanna,” and a little later I could get my answer. And, there’s always Highwire at Stanford to search for review articles on warfarin sensitivity.
Or I could, after all, actually talk to my doctor. Fiddling with my own warfarin dose isn’t something I’d be inclined to do, even with advice from my mother or the hope of a phone call from my sister to Joanna. More likely, I’d tell my doctor about the test result, and my doctor would decide whether to confirm it with a test from another lab.
But understanding genetic information on your own can be hard, and 23andme forums are full of confused people. Razib Khan, who is on the side of leaving 23andme alone, has nevertheless remarked on the number of posts he’s seen from people who didn’t understand that their low risk result still left them with a pretty good likelihood of developing common diseases. (Do not take a low risk result on 23andme’s breast cancer risk assessment as all that reassuring if you have a family history of the disease. Possibly because Myriad until recently had a patent on most BRAC variants, 23andme only tests a couple of them.) Conversely, just yesterday I saw a worried forum post about 23andme’s report on multi-proliferative neoplasms.
“I have a sister with Multiple Myeloma and I have Substantially higher odds of developing Jak2 V617F-positive MPN….”
I have a father who died of Multiple Myeloma and also got a 23andme result showing substantially higher odds of developing MPN related to a JAK2 variant. But it turns out that
- This condition is so rare that even my elevated risk is still tiny.
- 23andme is reporting, here, by their own account, on preliminary research.
- There’s no evidence that the gene variant on which they’re reporting has anything to do with multiple myeloma.
(Are you sure, Lynn? Don’t both you and this other 23andme customer both have the variant and have close relatives with multiple myeloma? Well, yes, I’m sure. I might have an as yet undetermined genetic factor predisposing me to multiple myeloma, though the genetic counselor I spoke to after my cancer diagnosis didn’t seem worried. But the fact that more than one 23andme customer both got this test result and had close relatives with multiple myeloma means little, since 23andme has many, many customers. I’ll wait for an actual study showing an association between this gene and multiple myeloma before I worry.)
Are there adjustments that 23andme could make to their disclaimers that would help in these situations? I don’t know. The disclaimers are already there, all over the site. If anything, the problem may be finding them among all the other stuff on the site. And I’m no expert on making medical disclaimers user friendly.
I do know that, however the current dispute between 23andme and FDA resolves, it’s just one step in a longer story of how we’re all going to handle this new technology. What kinds of regulation do we actually want for such tests? What’s reasonable consumer protection, and what’s unreasonable overregulation of innovation? And how will our lives be enhanced, or will our lives be the opposite of enhanced, by what we learn about our genetics?