You may have seen in the general press that the gene-edited twin girls born in China last year may have had their life expectancies shortened in the bargain. The doctor who edited the babies’ genes specifically edited one gene, that is associated with susceptibility to HIV infection. Their father is HIV positive, but that does not put the babies at any health risk. Further, the gene editing potentially could have increased their future risk for other infections. Now, a group in the United Kingdom have analyzed mortality data for about 400,000 people who volunteered to have their genetic information placed in a data bank. They reported that people who have a gene mutation similar—but apparently not identical—to the change made in the Chinese babies had a 21% lower chance of living to age 76 than people without the mutation. Now, the average age of the people who volunteered their information for the data bank is said to be 56.5 years, so the implication is that there is a shortening of life expectancy after middle age, for people who have lived at least that long.
One should interpret the U.K. analysis with caution, but the argument seems to be, “see, we don’t know the risks of human gene editing, so we shouldn’t be doing it.” And indeed we do not know the risks. But the argument in fact is, “…we shoudn’t be doing this—at least not yet.” As Joy Riley pointed out on this blog a few days ago, the goal of a moratorium on human genome editing appears to be to allow the scientists working on the technology to take time to build public trust and consensus for it. “We shouldn’t be doing this, ever” does not appear to be an option. Long-term readers of this blog may recall numerous posts over the last few years describing this process of gradual acceptance in the scientific community. The scientists draw an analogy to the 1975 Asilomar conference on recombinant DNA work, which established guard rails around that work. But the analogy is flawed. The risks of the work addressed at Asilomar were more readily defined, with shorter time frames to results, than can be addressed with genome editing. 400,000 middle-aged people’s mortality due to any (unspecified) cause over the ensuing quarter-century? How many edited people, studied for how long, over how many generations, with what consent process, to determine the risks? There can be no acceptable definition of the risks prior to actually assuming them. “The babies are the experiment.”
The correct conceptual framework for human genome editing is not benefit-risk analysis, it is deeper reflection on where we should not let engineering encroach on the human organism. “Keep your ambition off our bodies,” I suppose. And when we think in those terms, we should quickly recognize territory where we fear to tread at all, not just slow down.