U.S. releases the gene genie from the bottle

The U.S. Food and Drug Administration announced on Tuesday that the California personal mail-order genomics service 23andMe will be allowed to explicitly tell customers whether they have one of three common “BRCA” gene mutations that increase the risk of breast and ovarian cancer. It is difficult not to see this as a small but significant victory for regulatory sanity.

23andMe, named for the 23 chromosome pairs in the normal human cell, supplies customers who send in a saliva sample with the complete raw data on their genetic sequence. It is the same information that a high-paid genetic counsellor would use to derive and offer informed advice. But in 2013, the FDA decided that 23andMe’s “direct-to-consumer” marketing — letting people have relatively inexpensive access to the gene — was a problem. At that time, 23andMe was explicitly flagging gene variants that might have risk consequences in the materials it sent to buyers; the FDA determined that this amounted to the illicit sale of a “medical device.”

The company was ordered to stop calling attention to the dodgy parts of the gene, but it continued to sell the raw data to interested hobbyists, who could, if they were technically sophisticated enough to wrangle it, find out pretty easily what 23andMe wasn’t allowed to say out loud about their own cells. Most individual mutations, as 23andMe customers will know, are not especially consequential. But there are three mutations found at sites labelled BRCA1 and BRCA2 that are especially common among people of Ashkenazi Jewish descent, and unusually dangerous: the breast cancer risk for carriers is elevated about fivefold, and for ovarian cancer the factor is somewhere from 10 to 30.

This does not warrant an immediate mastectomy for women who test positive for the bad genes, which is a fear that the FDA came close to expressing when it moved against 23andMe in 2013. But BRCA status is information that women a) should have if obtaining it is affordable, b) should bring to the attention of their own physicians, who, in 2018, will be abundantly familiar with BRCA mutations and their consequences, and c) almost certainly warrants a level of close lifelong cancer monitoring that would be useless (if not positively harmful) for women in the general population.

The FDA is, of course, a trade regulator above all, and this makes it inevitably a defender of the commercial status quo. Its real impulse, no doubt, was an instinctive dislike for the idea of a bunch of Silicon Valley barbarians selling a sequencing product as a BRCA test per se, thus circumventing a laboriously built guild structure of physician knowledge and expensive genetic counselling (much of it done by licensed non-physicians).

What has changed in the past few years is that the FDA has realized that most of this is bound to happen anyway — in particular, that products like 23andMe, which have strictly recreational or genealogical value and which are offered by shops full of credentialed geneticists, are bound to become convenient venues for consumer genetic screening in the end. This is a historical development, an ongoing story, with particular relevance for everyone whose family tree features “founder effects” — i.e., a background of restricted genetic diversity owing to a period of social or colonial isolation. Jews of East European descent are just one example: if you are ethnically French-Canadian, or an old-stock Newfoundlander, or a descendant of Mennonites or Hutterites, you check this box.

When the FDA lost its temper with 23andMe, the company was careful — more careful, perhaps, than its Silicon Valley neighbours sometimes are — not to enter a violent publicity war with the regulator. The 23andMe people might have taken the logical position that they weren’t selling a “medical device,” since they weren’t and don’t offer customers any devices, other than a little plastic jar for your spit that you fill and send back to them.

But the company knew that the FDA had a “de novo pre-market review pathway” open to low-risk “devices,” and so it played ball. This “pathway” was itself opened up in the 1990s, as part of a culture change in U.S. medical regulation: by then it had begun to dawn on the establishment that being too conservative about new technology can be as harmful as letting it run riot in the marketplace.

The firm had to show that its genetic testing was as accurate as other forms on the market, which was fairly obvious (everybody from hospitals to crime labs uses pretty much the same offthe-shelf technology for the sequencing). It had to demonstrate that it could offer risk reports to customers that were as thorough and comprehensible as what a genetic counselling service might provide, which was not especially difficult in principle. The phrase “bureaucratic hoops” may be hovering in your mind right about now: that’s what they were. But — to the credit of the U.S. government — they weren’t burning hoops two inches across, of the sort regulators sometimes like to erect.