Amid ethics outcry, should journals publish the ‘CRISPR babies’ paper?

Like researchers everywhere, He Jiankui — the scientist in China who claims to have used CRISPR to edit embryos to create babies protected from HIV — is eager to publish scientific papers. It is, after all, a publish-or-perish world — although in He’s case, his fate at home may rest more with what the Chinese government thinks of his behavior than what a peer reviewer says about his work.

As STAT reported Monday, He shopped around a manuscript earlier this fall about using CRISPR to edit genes for a different purpose — to prevent an inherited condition that causes sky-high cholesterol levels — but it was rejected because of ethical and scientific shortcomings. And two weeks ago, in the face of withering criticism over his lack of transparency, He told the International Summit on Human Genome Editing in Hong Kong that he had submitted a paper on the “CRISPR babies” work to a journal.

Given the maelstrom surrounding He’s claims, however, should any journals even consider papers from him? And if they do, what should they keep in mind?

Jeremy Berg, editor of Science, told STAT that while he could not comment on whether the paper had been submitted to his journal, “given the numerous ethical issues with this situation as presented, we would be extremely unlikely to consider it.”

Howard Bauchner, the editor of the Journal of the American Medical Association, wouldn’t comment on the possibility of a submission by He either, but said, “I believe articles should be reviewed and not judged based upon what is written in the media.”

No editor likes retracting papers unless absolutely necessary, so it stands to reason that they might want to avoid papers that, once subject to scrutiny, will require the nuclear option in scientific corrections. So here are some considerations for editors — many of them requirements that journals created, but sometimes sidestep — should He’s paper land in their inbox.

First, there’s He’s own publication history — and lack of candor. As Art Caplan, one of the world’s most prominent bioethicists, noted in Medium, He “recently published an ethics paper (yes, an ethics paper!) called ‘Draft Ethical Principles for Therapeutic Assisted Reproductive Technologies.’” As mind-boggling as that now is in retrospect, here’s another twist: He and his co-authors did not disclose that He was working on such a project as he was opining on its ethics. “The paper should be withdrawn as fraudulent,” Caplan demanded.

The editor-in-chief of The CRISPR Journal, where the ethics paper was published, told STAT the journal is looking into whether the authors failed to disclose conflicts of interest. “They disclosed none but some are clearly in play,” said the editor, Rodolphe Barrangou of North Carolina State University. “I also am concerned about the obvious lack of transparency about the premature work.”

Caplan had been down this road before. In July, the Lancet retracted a paper he had co-written that set out some ethical guidelines for another controversial area of research — transplants of engineered organs — because the results it referred to had been found to have involved misconduct by the last author, Paolo Macchiarini. Caplan told Retraction Watch at the time that he and Macchiarini’s other co-authors — whom the journal had all cleared in the misconduct — had been duped by the now-infamous transplant surgeon. We’d suggest that journals considering He’s papers make sure to check whether any co-authors know exactly what He has been up to — much as the TSA asks whether anyone has given you any items to pack in your luggage.

Then there’s the issue of preregistration. For well over a decade, researchers have been required to post their plans for any human trial — with some small exceptions — on a publicly available database such as ClinicalTrials.gov if they want to publish the results. It’s not just a rule for journals; the Food and Drug Administration requires it, too. That’s supposed to happen before a study gets underway, although we’ve seen cases where it occurs after the fact, which is what happened in He’s case. His protocol wasn’t posted on the China Clinical Trials Registry until last month, long after the trial began.

Next, leading journals require that human research be approved by an institutional review board (IRB), or an equivalent. He says his study was approved by such a committee, but at a different hospital from the one at which he did his work. That’s a gray area, at the very least. And failure to obtain, or document, IRB approval is a legitimate reason for retraction — and not to publish in the first place — if only because it is easier to prove that than the fraud that may be lurking in a suspect paper.

If editors are satisfied that He has met those requirements, the larger ethical issues remain. The purported feat was “out of ethical bounds six ways to Sunday,” Caplan, of New York University, told STAT. It’s unclear whether the parents of these babies knew the risks of the procedure, and He’s rationale for it — preventing HIV — seems flimsy given how many other, and simpler, ways that can be accomplished.

Still, journals have certainly published studies that ethicists and others decry as unethical. In 2000, Marcia Angell, then-editor-in-chief of the New England Journal of Medicine, published one such study, about an HIV trial in Uganda, but criticized its ethics in an editorial in the same issue.

Although the editorial was a bold move, it is, of course, one way to have one’s cake and eat it too. Journals like attention, whether it’s from the press or from researchers who cite papers in their own work. Just look at how many editors trumpet their journals’ impact factors, a measure of how often papers are cited. The HIV-Uganda paper has been cited nearly 1,700 times. But credibility isn’t measured by a particular number, and justified criticism could lead to a reputational hit.

There are still questions about whether He’s claims are truthful — that his work led to the births of gene-edited twins. Of course, journals and peer reviewers can be fooled. It is extraordinarily rare for anyone to ask to review raw data, and even that can be faked. Some journals shrug and say science is built on trust, so it’s just fine to publish authors such as Macchiarini who have been shown to be real-life Pinocchios as long as the papers make it through peer review. We’d suggest that it’s a Bayesian world, and that it’s better to consider one’s priors.

And that suggests an argument that it would be better for He’s claims to be properly vetted, and that’s really only possible if they’re published, or at least posted somewhere. That’s the idea behind one proposal, by “Andy Biotech,” who describes himself on Twitter as a biotech investor. Let He post a paper as a preprint, which doesn’t imply any vetting in advance, nor stamp a journal’s imprimatur on it, and then let the world scrutinize it.

It remains to be seen if He is willing to go the preprint route; in Hong Kong, he told the summit he would not make the full results available for scrutiny before publication in a journal. Preprints are only now gaining steam in biomedicine, so we can’t know whether they would have helped scientists catch cases of fraud from recent history — those of Hwang Woo-suk or Haruko Obokata, for example. But He is certainly willing to take risks, so maybe he’ll change his mind and take this one. Or maybe journal editors will decide the risk is worth taking, and publish him.

The world is watching.

Sharon Begley contributed reporting.