DNA Dilemma: The Full Interview With the FDA on DTC Genetic Tests

In an interview yesterday, the FDA made some potentially controversial and confusing statements about consumer genetics, an area it's in the process of deciding how to regulate. I criticize the agency in my piece for not being fully transparent, so it's only fair that I live up to my own standards and post the entire interview here (I've edited out stammering and a few digressions for clarity, but all other quotes are faithful to the tape). Besides, since these issues can be very confusing, I wanted the FDA to have an opportunity to have all its statements on the record. Here is a transcript of the interview, with Alberto Gutierrez and Elizabeth Mansfield, respectively the FDA's director of the Office of In Vitro Diagnostic Device Evaluation and Safety and its director for personalized medicine in the Office of In Vitro Diagnostics in the Center for Devices and Radiological Health:

NEWSWEEK: During the recent congressional hearing, 23andMe's general counsel said that the FDA "encouraged them to proceed" freely in 2008. What specifically changed about their test panel between 2008 and now that caused the FDA to start to take some regulatory action? Alberto, I think you've touched on this a little with me before, that the concern was mutations with powerful effects on breast-cancer risk or drug metabolism—
tests that might influence medical decisions. Is that correct?

Alberto Gutierrez: They met with [former FDA commissioner Andrew von Eschenbach]. We were not there. We did meet with them on several occasions, but their claims of what they're offering have been changing constantly. What was there in 2008 I'm sure looks very different from what is there now. The drug-metabolizing claims were not there previously, and even some of the claims that directly touch on disease states, cardiovascular disease, breast cancer were actually not originally—they're much bigger now.

That's interesting. I think before when we talked, the cardio and breast-cancer stuff was not brought up. So even some of the common disease claims—what is the specific concern with them? My understanding was that a lot of those genes really don't nudge the risk up or down very much.
AG: The concern actually is with everything. The law requires us to clear devices or approve devices before they go into the marketplace when they make medical claims. The question with 23andMe has been whether their claims were medical claims or not. The original claims they were making really were very much on the edge. We actually told them that we thought they were medical claims, but it was at least possible that you could argue that they were not. Now clearly the claims they are making are medical claims. And to show you this is an issue we actually have been dealing with, we actually brought it up in SACGHS, that this was a point of discussion in 2009, because we kept hearing from 23andMe claiming that they were not making medical claims. We actually thought to have the Secretary's Advisory Committee at least state whether they considered the claims that they were making medical claims or not.

In our previous interview, and this is a direct quote, you said those health claims were "what we would consider to be moderate risk." I've recently seen [the FDA's director of the Center for Devices and Radiological Health] Jeffrey Shuren quoted as saying these claims are high risk, which I assume could mean stricter regulation. Which is it, moderate or high risk, and has something changed since our initial interview?
It always depends on what medications they can have, how they're offering it, we always look at what is the risk of an undetected false result, what would be the medical action taken, and whether there's high risk there or not. I don't think Dr. Shuren meant to imply that this would be [pre-market approval, or PMA], but that there might be some claims there that we may consider to be PMA-like, or PMA per encounter ...
Elizabeth Mansfield: In general, though, they have a variety of claims that span our normal risk classification. On a claim-by-claim basis, there are some that might be low, there are some that might be moderate, there are some that might be high. I think Dr. Shuren's assessment that these were "high risk" was as opposed to "no risk."

So the FDA would actually be going through these panels test by test, and looking at each one and seeing, Is this one low risk or moderate risk or high risk, and approving them specifically?
: Yes, in a general sense we would go claim by claim. Whether each claim would require a different submission or not—we'd be looking at each of their claims and looking to see that they would have data to credibly back the claims they are making.

So if one of these companies wanted to add a new test to its panel, it would have to get that specifically approved?
No. I want to make a distinction between what we would be doing now, and maybe what we would do in the future, and the reason here is that—Dr. Shuren put this on the table [unintelligible]—with these devices maybe it can collect a lot of data that may not be relevant at the moment. If later on data becomes available that shows that the company may be able to make a claim that doesn't exist now, is there a way for us to do this that wouldn't require necessarily a submission for each of them, or that we could bring together experts that could make the call that there is now enough evidence to make this kind of claim? That is something that we are discussing, that we are trying to figure out how we're going to deal with devices that generate a lot of data for which there is no use now but may be in the future, like with whole-genome sequencing …

[When we spoke last time], you had just sent five letters out and Illumina got one. Your concern in that letter, I believe, was that Illumina was selling a "research use only" chip and it was being used for nonresearch purposes by DTC firms. There were several reports I saw this morning indicating that the FDA has "requested" that Illumina stop selling chips to DTC genetic-test providers. Is that true in whole or in part?
We have asked Illumina to work with us to bring their devices into compliance with our regulations. We have not specifically requested that they stop selling them to anybody.

So they can continue supplying 23andMe and deCODEme with chips right now?
We're working with them on that …

Illumina obviously isn't the only chip maker out there. Why is it the only one that is getting these letters? Are the other ones going to be called in as well?
You can bet that with the original set of letters in 2009, Illumina was not the only one to receive a letter.

I'm sorry, I'm not familiar with the 2009 letters. What were those?
We sent letters to several companies with direct-to-consumer chip arrays to come in and discuss the issue.

And did those companies come in?
: Yes, and I think that's part of what you saw with the public meeting on June 30 on the array issue and why several companies want to figure out a way to [unintelligible] …

[The chips are labeled "research use only."] The DTC providers, when they sell these tests, aren't [necessarily] doing research. Doesn't that imply that Illumina does have to stop selling the chips to them?
If they continue to label them that way, yes, that would be something that we would most likely take a little further action on. At the moment, they're working with us and we will see what they can bring to us in a reasonable amount of time without completely blowing up their business or the market or anything …

So I guess if it's not being used for research only it needs a different type of [label or] approval?
That's correct…

Until that happens, though, is it correct that the Illumina chips will remain on the market for direct-to-consumer companies to buy?
[Pause] I think Illumina needs to figure out how they're going to move forward.

Okay … I want to move on to whether the issue with direct-to-consumer is actually providing data to people, or is it the interpretation algorithms these companies are using? So, would a company need to be approved just to provide a raw SNP list to people?
They would if they made medical claims about that data. If they don't make any medical claims about that data, then they're free to provide information as far as we're concerned.

What exactly would constitute a "medical claim?" Would pointing people to medical research papers [qualify]?
AG: It depends. There are rules as to how one can do that … Those rules are actually worked out pretty well, and they just would need to make sure they're staying within the rules.

Are those rules on the Web?
I don't know where the policy is. I would have to get it for you. It's an agencywide policy. I would have to find it for you. And it won't be that easy for people to follow it…

Alberto, I had heard a story, I believe it was in Rob Stein's article in [The Washington Post], about these women who were having prophylactic oophorectomies on the basis of direct-to-consumer genetic results that were "questionable."
That's a little bit taken out of context in the sense that it's not based on the direct-to-consumer [model]. The cases we heard of were based on lab-developed tests that were not properly validated. What I did say was that even though 23andMe doesn't make a direct claim as for ovarian cancer, they do provide information that links what they are giving the consumer to ovarian cancer.

So that's an example of where the interpretation is about pointing to a medical paper.
: It's more than that. They interpret the medical paper for them … Ovarian cancer is called the "silent killer." People actually don't know they have it until it's too late. And women that are high risk a lot of times choose to have their ovaries removed because they are afraid that it may not be caught in time. There's no diagnostic test for it. So the tests that we have seen—there was one that was taken off the market because of FDA action … We knew that in those cases most women who were being operated on actually were likely to be negative for ovarian cancer … The one we heard most about originally, it was a test called OvaCheck that was going to go to the market, and we believed it was not well characterized and it was going to create problems. It never made it to the market, and then in 2008 another one made it to the market called OvaSure, and the FDA wrote LabCorp a warning letter …

Thank you, that's very helpful to clarify. In general, when the FDA makes decisions about potentially fairly serious new regulation of a device or a test or a drug for that matter, does it do any sort of analysis and take into account the costs that regulation may impose on companies?
This is a more complicated question than you actually think. We have rules we're supposed to follow that we're supposed to apply evenhandedly to everybody about what exactly is a new product. The law requires new medical devices to come into the FDA to get approval …That's what the law is, so we don't actually get to make a decision that's based on the economic issues for that new device.

I'm a little bit confused. Can we talk about that in the specific context of direct-to-consumer tests? What kind of thinking was there about the costs?
[Direct-to-consumer tests] are a new medical device, and under the rules of the 1976 law, new medical devices require FDA clearance or approval before they go onto the market.

Right. But as part of that deliberation process, when you guys are considering whether to approve a device or not, do you take into account the cost that would impose on companies or the general impact on the industry?
No. No. Our review does not, no, we don't take into account cost.
EM: And Mary, that cuts agencywide. That is not considered in any of our reviews.

Okay. Thank you. I thought I had seen some obscure law saying there was a
requirement to take into [account] costs … Yes. Here it is. This is actually on the FDA Web site: "As part of its mission to supply economic analysis to decision makers, the Economics Staff in the Office of Planning conducts economic analyses of all important proposed and final regulations issued by the Food and Drug Administration. Each economic analysis includes an assessment of the costs, benefits, and cost-effectiveness of the action." So this is right on the Web. How can you not take—
So that is just, if you look at the first part, that is for all new guidances, for new rules and regulations.
EM: Which this is not.
AG: This is neither a rule nor a regulation.

I see. You're saying that because this is just like in 1976, these are medical devices, therefore they need approval— you don't consider this new regulation, it's just that the old regulations need to apply to these companies?
That's correct.

Okay, thank you. One more question: … We've had the hearing and there were obviously some fairly troubling things that came out of the GAO report. I'm wondering what impact the hearing is having on the FDA right now. Are you guys taking those findings into account at all? Is that influencing your decisions?
The findings of the GAO?

And other things that came out of the hearing.
No, we believed that these devices [should be] under regulation. It hasn't influenced us. Dr. Shuren did say that we should have been moving faster, and that we need to make these companies come in and get cleared for moving forward and selling their tests on the market. We will proceed with due diligence and try to help the companies come into compliance.
EM: The GAO report clearly did not undermine our feelings that these tests could present some risks to patients.

[Given] Dr. Shuren's comments that there should have been faster movement, are you guys feeling pressured to move fast right now?
: I think we are doing what we are supposed to be doing. We are obviously trying to get the companies to come into compliance and working with them to do so.

Note: After this story was published, FDA press officer Erica Jefferson said that she, and not Elizabeth Mansfield, made the following comment: "And Mary, that cuts agency-wide. That is not considered in any of our reviews." Jefferson made no other comments during the interview.