The apocalypse as a rhetorical device in the influenza virus gain-of-function debate
Arturo Casadevall, Don Howard, Michael J. Imperiale
The authors note that the “central nugget in the controversy is a disagreement on the risks and benefits of such experiments”. It is fair to say that nobody has squarely addressed this duo, so there is agreement. This is quickly morphed into a dichotomy whereby pro-GOF proponents stress the benefits and the anti-GOF group concentrate on the risks. While risk is an important issue it is not the only one. Some opponents of the work have focused on the virology and concluded that the touted benefits were overblown. Others who have not taken explicit stands, or who cannot be considered opponents have argued that predicting a pandemic is extremely difficult, if not possible. For such a discussion, see:
- http://www.nejm.org/doi/full/10.1056/NEJMp1307009
- http://www.sciencemag.org/content/342/6156/310.2.long
- http://www.atsjournals.org/doi/abs/10.1164/rccm.201305-0914ED?url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org&rfr_dat=cr_pub%3Dpubmed&#.VEQ1I4dAyTc
- http://embomolmed.embopress.org/content/5/11/1637.long
Accordingly painting the anti-GOF group as risk averse is an example of rhetorical play.
“Anti-GOF” scientists do not use apocalyptic rhetoric. Indeed, they point out the risk to human health that could arise from a virus escaping or being released from a lab, for there is no such thing as zero risk. Yet it is a simple fact that influenza viruses do, occasionally, provoke pandemics. If the opponents of GOF influenza research had ignored this they could not have properly formulated the notion of risk. It is part and parcel of influenza A viruses that brings us close to pandemics.
If we were discussing GOF of bunyaviruses, for example, that have not yet provoked a human pandemic, use of the word pandemic could be construed as hype or scare mongering.
With this virus group in mind in mind starting from March 2009 there has been a recent outbreak of severe fever with thrombocytopenia syndrome (SFTS), in Central and NE China. The initial case fatality rate was 30%. Chinese researchers quickly identified a novel phlebovirus that they called SFTSV, a member of the bunyavirus group. (Yu et al., 2011)
Back to influenza; in lay language, at least for some, a pandemic is akin to the apocalypse. Indeed a New York Times Editorial ran with “An engineered Doomsday”
Other journalists have used Frankenvirus and there is no problem with that, except that these are words that scientists rarely use without thinking.
Recently we have seen three cases of symptomatic H10N8 infection in humans, one of which was in an immune compromised individual who eventually died. This led to four papers on the infections and at least three commentaries in the scientific literature along the lines of H10N8: the next pandemic? The references can easily be found by a PubMed search (human, H10N8).
Extrapolating from three cases to a potential pandemic is a little strong.
The arguments for doing GOF research are not as solid and reliable as the proponents would have us believe. This is an area where science and scientists can get to work.
The editorial of Casadevall et al., propagates a common misunderstanding. We read that “humans are notoriously bad at assessing risks and benefits, and … it is difficult to see how we are going to resolve this issue”.
- As the scientists cannot agree among themselves it is logical to conclude that there is no consensus as to the benefits.
- The same is not true regarding risk. The world insurance industry had a market size of $4.1 trillion in 2010 and is expected to rise to $5.1 trillion by 2015.
Lloyd’spandemic.insurance.impactsv2.pdf
Just because it is hard to forge a consensus as to the benefits of GOF flu research, does not mean that the same is true for risk. Linking the two is not scientific.
• Scientists are notoriously upbeat about their work and systematically underestimate risk (Van Noorden, 2013).
Logically the influenza virologists should open up and consult risk experts who have far more experience than they.
Van Noorden R. Safety survey reveals lab risks. Nature. 2013 Jan 3;493(7430):9-10. doi: 10.1038/493009a.
Verikios G, Sullivan M, Stojanovski P, Giesecke J, Woo G (2011). The global economic effects of pandemic influenza, General Paper No. G-224 Centre of Policy Studies and the Impact Project, Monash University.
Yu et al., Fever with thrombocytopenia associated with a novel bunyavirus in China. N Engl J Med. 2011 Apr 21;364(16):1523-32. doi: 10.1056/NEJMoa1010095. Epub 2011 Mar 16.
The first pair
“Valuing knowledge: a reply to the epistemological perspective on the value of gain-of-function experiments”
Nicholas Greig Evans
This paper is focused on an earlier paper from Arturo Casadevall, Don Howard, Michael J. Imperiale published in mBio in September 2014.
The language in Evans’ paper it not run of the mill for most virologists. It considers how we appreciate and value knowledge and how it enters into the bookmaking of risk-benefit analyses. Knowledge enters into the benefit side. This needs more discussion that hopefully should be forthcoming.
Reply to “Valuing knowledge: a reply to the epistemological perspective on the value of gain-of-function experiments”
Arturo Casadevall, Don Howard, and Michael J. Imperial
The authors rebut two points they identified in the paper by Evans. SWH don’t pretend to master all the arguments and certainly less the literature and should not comment.
However, they attribute intentions to scientists, and this is worth comment:
“However, we anticipate that, in such [risk-benefit] calculations, proponents of GOF experiments would attempt to maximize benefit but that opponents of GOF research would attempt to maximize risk, there being no objective standards for either. We note that such calculations are currently being proposed primarily by opponents of GOF experiments, presumably because they feel that they can make a stronger case for risk than benefit, but we doubt that these will be convincing or definitive. Hence, we do not believe that risk-benefit calculations would ever yield a number that would be accepted by all parties, and consequently, such calculations are of only limited benefit.”
- Again it sees the “other side” as only being interested in risk. They are only interested in maximizing their case and play on the fear of risk. Some have been studying infectious diseases and their agents for a lifetime. To accuse or reduce them to mere sabre rattling doesn’t rattle.
- As mentioned above, the scientific claims in favor of GOF influenza research have been shown to be wanting. The “other side” has been doing the hypothesis breaking that the influenza virologists should have done.
- It argues that opponents of GOF are only interested in scoring points and not interested in science or public health, to the point that taking part in any risk benefit analysis with them is not worthwhile in the present climate. Perusal of the list of scientists who wrote to President Barroso, or signed the Cambridge Working Group statement, suffices to push this to the side.
Second pair
Can limited scientific value of potential pandemic pathogen experiments justify the risks?
Marc Lipsitch
Like the paper by Evans, this is a comment on the same paper by Casadevall, Howard and Imperiale.
Lipsitch gets into the details of GOF experiments and makes a number of good points in simple language. He reminds us of the obvious limitations of animal models and that the interpretation of adaptive mutations can be modulated by the genetic backbone of the viral strain. In addition the risk associated with engineering influenza viruses must be tempered by the fact that they can, occasionally go pandemic humans. In short the extreme risk scenario is one of catastrophic risk. This leads to addressing the unusual ethical issue of putting large number of people at risk without their knowing, with informing them.
Reply to “Can limited scientific value of potential pandemic pathogen experiments justify the risks?”
Arturo Casadevall, Don Howard, Michael J. Imperiale
Casadevall et al., avoid Lipsitch’s caution over the use of animal models. They note that “the experiments on H5N1 transmissibility unequivocally established that H5N1 had the biological potential to become mammalian transmissible, and that observation was a qualitative, all-or-nothing, result for which there is no need for statistical analysis: it happened.”
What happened? Experimentally mutations were accumulated in H5N1 virus genomes allowing droplet transmission of the virus between ferrets. In short it is possible for an avian virus to morph into a droplet transmissible virus. Given that Spanish flu was derived from an avian virus we already knew that this was a possibility.
The opposite is also true: it was shown by Wilson Smith in the 1930s that a human H1N1 influenza A virus could be adapted to growth in the mouse brain. In short a human respiratory virus could be adapted to growth in a very different organ in another species. Agreed, inoculation by man was necessary.
This dovetails into the issue of relevance discussed by both parties. In addition to knowing that “it happened”, a.k.a. it is a possibility, a decision maker also needs an idea of the probability of its occurrence. Here the avian influenza GOF experiments are silent. The numbers of experiments are so small and the range of strains tested that a conclusion is not possible.
Note that the avian influenza A H5N1 strains used by Chen, Kawaoka and Fouchier, were first isolated in 2001, 2004 and 2005. When the HA mutations identified by Fouchier were transposed into an Egyptian H5 backbone dating from 2010, the HA failed to bind to human upper respiratory tract tissue sections (Tharakaraman et al.).
Hence even if there was a degree of convergence between the mutations identified in the Fouchier and Kawaoka studies, it doesn’t tell us whether they are of relevance to circulating strains today, and therefore of use to a Health minister.
While avian influenza viruses probably assail humans more often than they might like to realize, resulting in both asymptomatic and symptomatic life threatening infections, a century of observation shows that only H1N1, H2N2 and H3N2 containing viruses have caused pandemics. This is the data.
Accordingly, we might conclude that the probabilities for H5N1 or H7N9 a novel human pandemic low. Perhaps. We have great difficulty predicting the future of virus evolution and influenza pandemics in particular. By contrast, tracking virus evolution is a totally different matter and is done all the time for many viruses. There is never any place for complacency when tracking rapidly changing human viruses and global influenza virus surveillance is imperative.
The point is to separate hypotheses and data and not takes hypotheses for more than what they are.
References
Tharakaraman, K. et al. (2013). Structural determinants for naturally evolving H5N1 hemagglutinin to switch its receptor specificity. Cell. 153, 1475-1485. doi: 10.1016/j.cell.2013.05.035.
Third pair
An avian H7N1 gain-of-function experiment of great concern
Simon Wain-Hobson
This discusses a paper reporting GOF experiment with an ostrich H7N1 influenza A virus.
Publication of this paper was accomplished by a commentary from two editors of the Journal of Virology and Tom Shenk of the ASM Journals Board explaining why the journal published it. See June 2014.
Briefly the commentary by SWH points out some weak spots in the virology which show why it will be very difficult to derive useful data from this study for decision makers, apart from the fact that, once again, it happened, an avian virus can be adapted to infect ferrets via the airborne route.
It emphasizes that the viruses did not loose pathogenic potential with the acquisition of droplet, or airborne transmission. Approximately 60% of animals died, or in other terms the ferret CFR = 60%. Whether or not this would be the same in humans, we will never know hopefully. By comparison, and with obvious reservations, it is worth mentioning that the CFR associated with Spanish flu was around 2%.
The piece then points out the contradictions between Dermody et al. and the ASM Statement on Dual Use Research of Concern and Biosafety. See July 31, 2014. Suffice to say that former write “The ASM editors did not reach consensus about whether the work represents DURC. However, it was concluded that the benefits of publishing the paper outweighed any potential risks, and ASM decided to move forward with publication”. We were not given any reasons for such an opinion, nor were any references given. By contrast the ASM statement asks the US NAS to help out with the risk-benefit analysis that now seems necessary for GOF research.
The decision to publish an avian H7N1 influenza virus gain-of-function experiment
Terence S. Dermody, Arturo Casadevall, Michael J. Imperiale, Rozanne M. Sandri-Goldin, Thomas Shenk
The editorial focuses itself on the editorial review process of Sutton et al. While the peer review process is an uncontested accepted part of science, publication constitutes the toughest peer review. Reproducibility and credibility among fellow researchers are the sought after hall-marks.
Concerning risk-benefit analyses “we do not think an accurate (i.e., quantitative) risk-benefit analysis can be performed when neither the risk nor the benefit can be measured in a meaningful way.”
As there is strong disagreement among scientists as to the merits of the work it makes that establishing benefit is fraught with difficulty. Not being able to quantitate risk is another matter. Perhaps the community should broaden out and seek advice from risk professionals.
Despite the fact that “Benefits are similarly difficult to quantify”, drug and vaccine benefits are mentioned. These points have been dealt with (Mahmoud, 2013; Wain-Hobson 2013; Wain-Hobson 2014)
References
Mahmoud, A. Gain-of-function research: unproven technique. Science. 2013 Oct 18;342(6156):310-1. doi: 10.1126/science.342.6156.310-b.
Wain-Hobson, S. Pandemic influenza viruses: Time to recognize our inability to predict the unpredictable and stop dangerous gain-of-function experiments. EMBO Mol Med. 2013 Nov;5(11):1637-41. doi: 10.1002/emmm.201303475. Epub 2013 Oct 24.
Wain-Hobson, S. The irrationality of GOF avian influenza research. Frontiers in Public Health 2, 77 2014 doi: 10.3389/fpubh.2014.00077
A friend permitted reproduction of several paragraphs from a closely argued email:
“While the ASM DURC Review Committee correctly obtained opinions from the University of Maryland Institutional Biosafety Committee (the IBC for the host) and NIAID (the funding agency), this was not sufficient because of the inherent conflicts of interest and because the IBC addresses biosafety, not biosecurity, and there was no claim that biosafety was an issue regarding the H7N1 research.
With reference to the question “whether the resulting knowledge, information, products, or technologies” met the definition of DURC as defined in U.S. policy, according to the ASM, this requires an interpretation of what precisely a criminal needs to know to perpetrate a crime. This is a judgment call for which reasonable people may not come to consensus. Therefore, it should not be surprising that individuals and institutions can disagree about whether a particular study constitutes DURC.
But from where came the concept of “crime”? The DURC analysis is not one of criminal law but whether the information can be “directly misapplied to pose a significant threat with broad potential consequences to public health and safety, agricultural crops and other plants, animals, the environment, materiel, or national security.”
Additionally, although the ASM editors rightly insist that “the categorization of scientific work as DURC does not preclude publication,” it does not follow that “the decision to publish such work should be made independently of whether a study is considered DURC.” To the contrary, DURC remains one of several key factors in the decision, no matter how difficult that assessment may be or how wrong it is that the burden falls on journal editors. Further, it is a mistake of the ASM editors to rely on the NSABB’s recommendation about the two H5N1 papers as either precedent or indicative of “contemporary practice.” First, a joint publication of two papers hardly constitutes a precedent or any significance, let alone an indication of contemporary practice. Second, the NSABB’s first recommendation was that neither paper be published. Both manuscripts then were revised, so it also is erroneous that the ASM claims that they were “unredacted.” Then, although the second recommendation about the revised Kawaoka paper was unanimous that it be published, one third of the NSABB members disagreed with the recommendation to publish the revised Fouchier paper.”
