<p>Calmom - the problem with 23&me is that they want to have their cake and eat it too. They want to make all kinds of extravagant medical claims for their tests’ ability to predict risk of disease without having to go to the trouble and expense of demonstrating to anyone that it is actually true.</p>
<p>You apparently accept that the company’s tests can accurately assess your risk of of all those diseases simply because the company said so. Having dealt with many substandard lab tests over the years that range from the sloppy to the outright fraudulent, the FDA is far less trusting. Their attitude is more one of: “In God we trust. All others must show valid data.” </p>
<p>Other more ethical companies do the necessary studies and file for regulatory approvals all the time; hundreds are filed and approved every year. 23&me’s business model apparently assumes that the law doesn’t apply them, that FDA regulations are for chumps and not for cool and intrepid entrepreneurs like themselves. Well, the law says otherwise. The FDA has tried to work with them to get them into compliance, so far with no success. Now the FDA is losing patience with their continued flouting of the law.</p>
<p>This has me curious for my kids who have Gentile and Jewish heritage on both sides. My maternal grandma jokingly called herself a “Frew” (French Jewish). H’s dad was German Jewish. My Italian dad’s ancestry included Sicilian Jewish on his father’s side. We don’t know if the Jewish heritage is Ashkenazi, Sephardic, both or neither. Italian/Sicilian grandfather died of heart issues. </p>
<p>We don’t know much detail about the Gentile ancestry either. Paternal grandma (Italian/French) died of stomach cancer. Maternal grandfather (French) died from lung cancer, but he was a chain-smoker so don’t think that’s genetic. My mom died from strokes. My dad was 90 when he passed, but he did have heart issues and prostate cancer. H’s mom (Irish) is suffering from strokes. H’s dad died from heart and liver issues (non-alcohol related).</p>
<p>We don’t appear to have breast cancer in our family, but we do have heart issues and strokes on my side. H’s side also has some heart issues (valve), but lots of rheumatoid arthritis and thyroid issues.</p>
What “extravagant” claims-- the report is very clear as to the basis of the data. </p>
<p>
</p>
<p>But they didn’t say so. They use the word “approximate”, they have a disclaimer on every page saying the results are for educational and research use and not diagnostic, AND they tell me the strength of the research, citing to specific studies.</p>
<p>
</p>
<p>If you read the article at the link I provided, you’d know that this is incorrect. There are different levels of approval required for different types of uses/devices. It looks like 23andme has been filing 510(k) letters with the FDA, but now the FDA wants more:</p>
<p>
</p>
<p>In other words, if the FDA ruling stands and is applied to other genetic testing labs, your doctor won’t be able to order up most tests either</p>
<p>I’ve just done a little more research (of the legal type) and have confirmed that an FDA warning letter (which is what was sent to 23andme) is not legally binding. That is, in theory 23andme could simply ignore it and go right on selling their product – the FDA letter is the legal equivalent of the sort of cease and desist letters lawyers send out before commencing a civil suit. In practice, a company which has received a warning letter will submit a response and perhaps suggested corrective action within 15 days of receiving the letter. That’s this part of the letter:</p>
<p>
</p>
<p>So basically, the legal status is that 23andme is likely to continue selling its kits, but will have its lawyers write a letter to FDA listing the corrective action it plans to take.</p>
<p>One thing you have to remember about FDA is that they carry a badge. They are not merely attorneys - they also have law enforcement authority. And the FDA is not bluffing here. In the warning letter they are calling the product “misbranded” and “adulterated.” In FDA-speak those are serious fighting words - about one or maybe two steps away from filing criminal proceedings against the company and its officers if they fail to shape-up and get serious about following the law. </p>
<p>And when the FDA ordered the company to to stop “marketing” the test, they didn’t mean merely to stop advertising it. They mean stop selling it. If the 23&me is stupid enough to continuing selling their stuff in the face of this warning letter they’ll deserve every bit of the legal punishment they’re going to get.</p>
<p>“In practice, a company which has received a warning letter will submit a response and perhaps suggested corrective action within 15 days of receiving the letter.” Exactly, and the corrective action that was suggested is “…immediately discontinue marketing the PGS until such time as it receives FDA marketing authorization for the device.” Coureur is correct - in FDA speak it means “stop selling this test”, and they have the means to shut down the company if it does not comply.
[Office</a> of Criminal Investigations - Wikipedia, the free encyclopedia](<a href=“http://en.wikipedia.org/wiki/Office_of_Criminal_Investigations]Office”>Office of Criminal Investigations - Wikipedia)</p>
<p>It’s a warning letter, not an order. The meat of the letter is that it directs 23andme to submit a response within 15 business days (essentially 3 weeks from the date of the letter). </p>
<p>I know that it is worded in a way that it sounds like an order to a lay reader, but a warning letter doesn’t have the force of an court order or injunction. (And it isn’t worded as an “order” in any event – the “Therefore…” paragraph is a statement of the FDA’s legal position. The next paragraph after that tells 23andme what they have to do: provide a written response that details corrective action. </p>
<p>I’m just looking this from a legal point of view. I think it’s possible that 23andme came to an impasse in negotiations with the FDA, and wants to invite court action-- that is, maybe their position is that the kit is not a class III device, but rather an exempt device - see [Medical</a> Device Exemptions 510(k) and GMP Requirements](<a href=“http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpcd/315.cfm]Medical”>Medical Device Exemptions 510(k) and GMP Requirements) – if so, they might send a letter to the FDA laying out their position. The FDA would have to respond by bringing court action-- and maybe 23andme wants that. </p>
<p>The other possibility is that they simply dropped the ball – their general counsel who was in charge of working with the FDA left in July, so maybe 23andme just wants to pick up where they left off in prior negotiations. </p>
<p>But it is obvious that 23andme has received the letter and has not stopped marketing the kit. The website is up, the shopping cart works. (It would take about 10 minutes to post a notice on a web site and shut down the ability to place an order – they didn’t do that.)</p>
<p>I seriously doubt that the FDA does its negotiations in the courtroom. If you think that the reasons they issued the warning letter will be settled by a judge or jury, so be it. I happen to think that the FDA is prosecutor, judge, and jury here. The company either complies or does not at the risk of civil and criminal penalties. </p>
<p>The fifteen days are mere boilerplate words. This company is de facto shut down as the remedies are simply beyond their reach. The FDA wants them to obtain an authorization from … the same office that says 23andme has not been in compliance despite the previous requests. The website might be up because of the holidays and the patience of the FDA for a mere couple of weeks. The warning IS a cease and desist. It is an order. Not a slap on the wrist.</p>
<p>Given the deep pockets behind the company, there are plenty of reasons to believe this company will NOT continue this saga if financial or criminal penalties loom large. You can safely expect a response on the website that will amount to a “Sorry we are on a forced hiatus” </p>
<p>They probably will not be luckier under different regulatory bodies. Frankly, it does NOT look good for them as the chances to obtain an approval based on their possible claims is slim to none. And they probably have known this for a long time considering the frivolity of their previous applications.</p>
<p>Bunsen, I usually don’t rely on wikipedia for my legal research. </p>
<p>The warning letter is not the start of a criminal investigation. It is the start of a civil process that typically goes like this:</p>
<ol>
<li>FDA sends warning letter, giving company 15 days to respond</li>
<li>Company sends response</li>
<li>If FDA doesn’t think response is adequate, they send a subsequent warning letter, giving company 15 days to respond</li>
</ol>
<p>This process can go on for months or years. </p>
<p>When a case is resolved it will the database will show a “close out” date" – there is a huge number of warning letters from past years that have no close out date without much action taking place in the interim. </p>
<p>The 23andme response probably will be posted to the web site-- so we’ll see when they respond what the next step in. </p>
<p>The only criminal penalties are a misdemeanor – $1000 fine - so unlikely to be utilized in a case like this. The civil penalties have a little more teeth, though not enough to deter a company with mega-bucks – but the point of any legal enforcement would be to enjoin the sale of the product. See [21</a> USC § 333 - Penalties | Title 21 - Food and Drugs | U.S. Code | LII / Legal Information Institute](<a href=“http://www.law.cornell.edu/uscode/text/21/333]21”>21 U.S. Code § 333 - Penalties | U.S. Code | US Law | LII / Legal Information Institute) for penalty information.</p>
<p>“This process can go on for months or years.”</p>
<p>Certainly. The FDA letter sounds like it was written by a person who is losing patience (the dramatic preventive mastectomy reference, which is plausible but highly unlikely - I would not have included such a drastic example of what might happen). The 23andme saga has been going on for at least 3 years, and it looks like the company has not made any progress towards getting the information requested. Instead, they added indications to their laundry list of medical conditions and beefed up their marketing campaign. Of course, no one is going to jail here and no criminal investigation is needed - that would be ridiculous. What the FDA can do is to get a court order to stop sales of their product with the current labeling if the company refuses to do it voluntarily - until the product is fully compliant with the FDA regulations. What the company can and should do is to continue selling their genealogy kit, which is not designed to diagnose medical conditions. Also, it sounds as if they are in need of a qualified general counsel and might even have a position open - go for it, calmom. ;)</p>
<p>A side note, folks with Google-deep pockets should not cry about the costs of PMAs. Many much smaller companies end up having to file them and get them approved. Can’t wait to see what anti-aging drug Calico will bring to market without any clinical trials ;)</p>
It’s possible that they don’t see eye to eye about what sort of information ought to be necessary. 23andme doesn’t manufacture the testing array – rather, they are using a product made by Illumina. The FDA has just approved several Illumina DNA tests – so apparently Illumina knows what is doing to get approval. I would also assume that 23andme is one of Illumina’s biggest customers… so I would expect they have Illumina on their side. </p>
<p>I still sense that there is stuff going on behind the scenes, and I really don’t know what it is. The 23andme reaction to the warning letter was very muted – I have a hard time believing they were caught by surprise on this one.</p>
<p>Illumina sells its sequencing instruments and kits to anyone willing to pay the hefty price - what the buyer does with the instrument is not a concern to Illumina. Selling scientific instruments is not as profitable as selling FDA-approved diagnostics. Note that Illumina did get the approval for both the platform (MiSeqDx instrument) and a couple of diagnostic tests. These approvals were highly hyped up by the analysts covering the company, and I agree that it is a huge deal. Would Illumina turn into a diagnostic company or would it continue to sell its platform to anyone wiling to develop their own tests and pay Illumina hefty license fees, royalties and reagent fees? Selling scientific instruments is not as profitable as selling FDA-approved diagnostics, so… I highly doubt that Illumina is going to “help” 23andme deal with their FDA issues (I think you understand that their product as it is currently being marketed will never pass the FDA’s requirements). It will be in the Illumina’s best interests if 23andme simply re-labels their products so they would not carry any suggestions of having a diagnostic value and continues to buy RUO reagents from Illumina.</p>
<p>I work in the biomedical field, and over the years I have personally been involved in conducting the necessary clinical trials and the filing of many successful 510k(k) and PMA submissions with the FDA. So I know only too well the expense and effort all this takes. </p>
<p>I also unfortunately know what means when a company gets an FDA warning letter. I’m sorry to that we got one once, and it is very much a Big Deal - not some mere bureaucratic formality or regulatory bluff. It means the company is in a lot of trouble, and the only way they are going to get out of this trouble any time soon is by rolling over and starting to do everything the FDA demands. </p>
<p>I can’t imagine that anyone who knows what they are doing would continue selling a test after they have been told in a warning letter to stop. If anyone there is even remotely dreaming of slapping an already-angry FDA in the face like that, they should immediately hire a good law firm out of DC that specializes in FDA regulatory cases, and then start following that law firm’s advice. </p>
<p>At this level of violation the officers of the company won’t go to jail unless FDA turns up evidence that someone was seriously harmed based on the use of their test. But they may well lose their jobs. The FDA may well require a clean sweep of the responsible company officers as part of the clean-up effort. </p>
<p>Illumina is located right here in San Diego, and I happen to personally know a few of the Regulatory people over there. An as far as I can tell they are stand-up, ethical folks who know the rules and play by them. Apparently the same cannot be said about their customer, 23&me.</p>
<p>With Google money behind them, the thing that to me rings most false about 23&me’s arguments is that they don’t have the resources to conduct PMA-level clinical trials on their test. Nonsense. Companies do these kinds of clinical trials all the time. Are they time-consuming and expensive? Sure they are, especially when compared with what feeble efforts 23&me has been making up till now. But they are completely doable. Companies with far less resources than theirs do them successfully every year.</p>
<p>And the more I think about it the more I conclude that coming from Google may be at the root of 23&me’s problem. They are coming from a completely different business model, mindset, and regulatory landscape. They moved from the barely-regulated wild west of silicon valley computer-land into the highly-regulated world human health care, apparently without understanding what that really means. And now they are finding out the hard way.</p>
<p>There is indeed a world of difference between 23andme and companies such as Illumina or it’s competitor Life Technologies/Fisher. The latter having been the target of multi-billion acquisition offers. The products and devices might very well be a 100 billion market. There is no comparison with a company selling a 99 dollars quick test (or the expanded genome at 1000 bucks.)</p>
<p>It is remarkable that the FDA issued the warning letter THREE days after approving Illumina’s new device and tests. It should be clear that the FDA has lost patience with the companies that are attempting to sell to the masses by circumventing traditional channels. The companies that were targeted in 2010 have rolled over and bitten the dust. </p>
<p>It is also remarkable that 23andme has appointed a CEO from an industry totally foreign to the medical field. This seems to indicate that they are entirely oblivious to the depth of their problems and probably massively ignorant about the precise science they are peddling. </p>
<p>The only redeeming value, in addition to the spiffy and entertaining ancestry tool is that it might have generated a lot of interest and potential demand for a genome or exome test to be administered by a health service provider at a VERY affordable cost. It is not far farfetched to expect more complete tests at the cost of full blood tests or PCR analyses in the next years. </p>
<p>But the 99 dollar version will NOT be in your Xmas tree in 2013.</p>
<p>Coureur, all I am saying is that you can go to 23andme and see that the product is still being sold. I don’t work in the biomedical field and I am not denying that most companies would see the letter as a BFD, but I was once a lawyer and I am now a web techie, and I know that it would take no time at all to post a notice on the web site that says that the product is temporarily unavailable – and they haven’t done that. </p>
<p>I’m confused by your statement about clinical trials. Where do you think a clinical trial would be appropriate? Do you think that 23andme needs to prove to the FDA that the Illumina panel they use is accurate? 23andme doesn’t make a medical diagnosis – it just tells people their genotype for specific factors and then gives a summary and link to previously reported medical journal articles. I’m at a loss to figure out what sort of “clinical trial” would be expected in this context.</p>
<p>I mean, when I think of a clinical trial, I think of the context of a double blind pharmaceutical test-- but 23andme isn’t prescribing drugs or making any medical recommendations beyond “see your doctor” or a whole lot of common sense health advice. </p>
<p>I don’t know what the FDA wants from them or what they have already given - I’m just trying to get a sense of what they could give relevant to the service they are offering – which is to do a genetic test and explain what it means. That is, the only concerns that I have as a consumer is that the testing is accurate within whatever is expected for the industry --(even the best labs in the world sometimes make errors) – but as they are using an outside lab it seems to me that would be a matter of that lab’s quality control protocols. </p>
<p>So, if we assume for a moment that 23andme has already delivered to the FDA a detailed description of its sample handling procedures and protocols, together with a spreadsheet which provides a list of all SNP’s that the Illumina test identifies correlated with the bibliographic references that 23andme has accumulated for each SNP … what more is needed? and why?</p>
<p>In other words: 23andme tells me that my genotype for the marker rs4988235 is AG. They tell me that the A version is associated with lactase production, so my genotype means that I am likely to be lactose tolerant – although they also state that I could be lactose intolerant for “other reasons”. They cite to and link to 5 published studies. This is reported to me on a page with a prominent notice that says, “The information on this page is intended for research and educational purposes only, and is not for diagnostic use.”</p>
<p>So taking that particular trait - what is 23andme supposed to do? Replicate the work in the 5 published studies? Conduct its own independent trials and surveys to determine how consistent the association of a GG genotype for rs4988235 is associated with lactose intolerance in a large group that is demographically similar to its customer database?</p>
<p>I’m not trying to give you a bad time – I’m just really bewildered about what the FDA thinks that 23andme should be doing beyond what it already does.</p>
<p>I think the short version is that the FDA wants to see a complete PMA. Regardless of the attempts to market it differently, the FDA considers the test sold and its interpretations to be well beyond a Class I type. Right now, they probably consider it a de facto Class III that they might reclassify as Class II upon review.</p>
<p>No, I don’t understand that. I am totally baffled at what the FDA requires. Is there a reason to doubt the accuracy of the data produced by the Illumina panel that 23andme uses? </p>
<p>Or does the problem get resolved simply by 23andme changing the language used on their web site and other marketing materials? That is, if they changed the language on their home page from “Reports on 240+ health conditions and traits” to, say, “Reports on more than 240 genetic markers” – would that be the type of change that would satisfy the FDA?</p>
<p>What you describe here actually exists as a legitimate regulatory pathway. Its common name in the industry is called filing a “paper PMA” - where you merely cite published scientific literature without conducting any actual clinical testing. The FDA is real proud of having established this, but nevertheless it very seldom attempted and very seldom successful when it is. Why? Because the published papers are usually discovery research studies that were never intended or designed to satisfy the requirements for FDA approval. Thus they are hardly ever done to the level of scientific, demographic, and especially statistical rigor that the FDA demands.</p>
<p>The research studies are usually based on a sample of convenience - the doc had some frozen DNA samples left over from some unrelated study, or he recruited say a few dozen medical students that were handy on campus. Thus they were all young, healthy, mostly white people from one small geographic region - hardly representative of the type of diversity of age, race, and health status that the 23&me test is intended to serve. And most importantly, the sample sizes are almost always way too small in those studies. And combining a bunch of small studies to add up to a large study usually won’t work either, because the rules of clinical research require that for results from different studies to be poolable they must have all been conducted under indentical or nearly-identical protocols, which almost never happens. </p>
<p>I’ll tackle your example of the lactose tolerance test in the next post.</p>
<p>This is the type of “extravagant medical claim” I was referring to back in post #81. They are telling you that you are likely to be lactose tolerant based on three different things that have not yet been demonstrated to the FDA:</p>
<ol>
<li> That the specimen they tested was collected, shipped, and stored in a way that preserves its integrity and stability and thus its ability to be the basis of generating correct information.</li>
<li> That the assay used to test it accurately measures the gene in question, free from the many possible sources of interference and analytical error.</li>
<li> That your results tested with 23&me’s assay actually means that you are in fact lactose tolerant. Not merely that some other quite different test for the same gene developed by some researcher at a medical school previously showed that association, but that 23&me’s test shows that association too.</li>
</ol>
<p>All three have some difficulty involved, but #3 is usually the toughest. This is where a paper PMA would be great, but like I said, it hardly ever is actually possible to do. And the reason why FDA just won’t accept a small paper or two in place of #3 is that the field is littered with examples where one test for a given target works great, but another one that purports to measure the same thing fails to show the same medical result.</p>
<p>So what to do? What sort of clinical trial could 23&me do to show that their result actually means that you are likely to be lactose tolerant?</p>
<p>Okay let’s design a simple clinical trial right here:
They go to three or four locations scattered across the country (chosen to provide geographic and racial diversity) and from each site they enroll a bunch of people in the trial and pay them say $25 to provide some saliva to be tested.</p>
<p>Those specimens are sent off to the company to be tested in their assay. Separate from that the enrolled people are tested by the docs to diagnose whether are or are not lactose tolerant - based not on the results of the 23&me test (that would circular) but on their symptoms and medical history. </p>
<p>The docs doing the diagnoses are blinded to test results and the people doing the lab tests are blinded to the diagnoses. When they have enrolled say 100 people with and 100 people without lactose tolerance (100 is just a round number I picked for this example), the blind is broken and we all look to see whether the people who have the form of the gene said to be associated with lactose intolerance are in fact more more likely to be lactose intolerant. If so, Yay! The test is demonstrated to be clinically valid, we write all that up and ship it off the FDA and our PMA gets approved. </p>
<p>And THEN 23&me can legitimately market the test for determining whether or not people or lactose in/tolerant.</p>