At the heart of the current debate over the origins of the COVID-19 pandemic is the question of how to manage the risk of future infectious disease outbreaks. For over three decades, experts in emerging viruses have argued that zoonotic spillover is an ongoing and inevitable process, and that human incursions into the natural environment have made the world increasingly susceptible to catastrophic pandemics. From this perspective, ever since the appearance and spread of HIV/AIDS, “the next pandemic” has been right around the corner.
However, the recent rise of an alternative theory of the pandemic’s origins poses a disturbing question: Is COVID-19 the realization of these experts’ prophecy? Or did the prophecy in fact bring the pandemic into being?
We Are Not Prepared
With a return to normalcy in sight (at least in some parts of the world), we are entering what might be called the “post-hoc assessment” phase of the pandemic. This phase typically involves a process of collective diagnosis, in which official commissions draw lessons from the event that point to the need to anticipate similar future crises. Thus, a World Health Organization review of the response to the 2009 H1N1 pandemic concluded that “the world is ill-prepared for a severe pandemic or for any similarly global, sustained and threatening public health emergency.” Five years later, after the catastrophic Ebola epidemic in West Africa, another W.H.O. review committee warned that “the world cannot afford another period of inaction until the next health crisis.” Such assessments, in turn, seek to galvanize resources and structure organizational reform in order to ward off future catastrophes.
Initial post-hoc assessments of the world’s response to SARS-CoV-2 share this basic structure. The pandemic was a “preventable disaster,” an independent panel of experts established by the W.H.O. recently reported, pointing to “weak links at every point in the chain of preparedness and response.” Having slept through prior warnings, the panel argued, the “world needs to wake up” in order to adequately “prepare for the future.” It recommended a number of reform measures, including an improved system of global disease surveillance and a platform for the rapid and equitable production of medical countermeasures.
But just as this process of post-hoc assessment was getting underway, an unsettling counter-diagnosis gained sudden and unexpected currency: the long-marginalized hypothesis that the pandemic originated not with a spillover event in the wild, but rather through the accidental release of an experimental virus. If validated, this counter-diagnosis would not only attribute blame very differently — it would also fundamentally transform how we evaluate future disease threats. It would point not to the familiar assessment that “we were not prepared,” but rather to the conclusion that — at least in one critical way — we were too much in thrall to calls for preparedness. Indeed, it would suggest that the very demand for preparedness may have sparked the catastrophe.
This interruption of the post-hoc assessment process is the result of a series of recent events: In mid-May, a number of highly regarded scientists published a letter in Science calling for further investigation of COVID-19’s origins; at around the same time, respected science journalists began to suggest that the “lab leak” theory should be revisited; and then the Wall Street Journal reported on U.S. intelligence findings that several scientists at the Wuhan Institute of Virology, where research on bat coronaviruses was taking place, were hospitalized with COVID-like symptoms in November 2019, before the first officially reported cases of the disease.
Meanwhile, scientists have been unable to identify an intermediary host animal that could confirm the spillover hypothesis, adding to the plausibility of an accidental laboratory release as the initial source of the outbreak. Thus, we find ourselves in a situation of diagnostic uncertainty, both about how to attribute blame and about the horizon of future reform.
Much of the public commentary on this newly credible origin story has emphasized what it would imply for culpability — whether of the Chinese government, experimental virologists or an overly credulous media. But beyond the search for a proximate culprit, the resurgence of the lab leak hypothesis points to a larger series of questions about the conditions of possibility for such an event. It leads us to ask about the surprising confluence of actors at the heart of the story of a possible laboratory accident: An environmental NGO based in New York devoted to collecting samples of viruses from wild animals around the world; a virology laboratory in Wuhan doing experimental research on bat coronaviruses; and the U.S. National Institutes of Health, which supported a collaboration between those two entities as part of a larger portfolio of funding for basic research in the virology of emerging diseases.
We must try to understand: What principle of intelligibility guided researchers to take blood, saliva and fecal samples from bats living in obscure caves in southern China, bring these samples to a laboratory in a faraway city, analyze viral fragments found in the samples and experiment on them to explore their transmissibility among humans? How, in other words, did disease ecology meet experimental virology under the auspices of government research to improve public health?
A good place to begin the story is in the late 1980s, at the height of the HIV/AIDS crisis and toward the end of the Cold War, as a group of life scientists began to consider the future of infectious disease. It is here that we can find an initial articulation of the idea that monitoring animal viruses in the present might curtail future epidemics in humans.
In May 1989, an interdisciplinary conference on the topic of “emerging viruses” was held in Washington, D.C. The premise of the conference was that public health must take into account the centrality of human action in driving the emergence and spread of novel infectious diseases. Participants pointed to examples such as the evolution of new strains of influenza as a result of farming practices in Asia that intermingled ducks and pigs, the spread of dengue across oceans via water lingering in used tires and the role of new techniques of maize production in extending the range of Argentine hemorrhagic fever.
One of the conference organizers, the virologist Stephen Morse, introduced the concept of “viral traffic” to describe the movement of novel or already-existing viruses to new species or new populations. As Morse described it, the emergence of new infectious diseases was a scientific story, but also a moral one — about the unintended consequences of modernization and the ecological degradation that accompanied it.
Thus, for example, large dam projects in the developing world had led to the expansion of mosquito-borne diseases such as Rift Valley fever. Industrialized meat production in the U.K. had fostered the emergence of mad cow disease. Large-scale rural-to-urban migration often introduced remote pathogens to larger populations, as in the case of the virus causing Lassa fever. Meanwhile, the advent of international air travel meant that an obscure disease could rapidly spread around the world.
The story of emerging diseases also pointed to the need for new forms of public health intervention. If we are to manage the threat posed by novel pathogens, Morse argued, we must become better viral “traffic engineers.” An initial step would be to establish a global disease surveillance mechanism: A network of monitoring stations located in tropical areas and staffed by trained field epidemiologists could alert international health authorities of the emergence of a novel and deadly virus. This surveillance network should be attentive to “viral traffic signals,” such as deforestation, dam construction, disruptive changes in agricultural practice or major population migrations.
Such anticipatory knowledge might curtail the onset of future pandemics: “Most viruses that today are worldwide were once localized,” Morse observed. AIDS, for instance, had begun as an emerging viral disease. If the right tools of detection and containment had been in place, it “could have been stopped at the pre-crisis stage.” But the goal of pandemic prevention was not yet technically feasible — it would require the development of a “methodology for assessing the likelihood that a given animal virus will emerge as a human pathogen.”
In the meantime, according to Morse, the ongoing emergence and spread of novel pathogens into human populations were inevitable. Recent epidemics “should force the realization that new viruses will always be imminent,” and that “tragedies like the AIDS epidemic will be repeated.” As industrial modernization caused further ecological degradation, “episodes of disease emergence are likely to become more frequent,” he predicted.
To address this intensifying threat, support for research into the process of disease emergence was needed. With a better scientific understanding of viral evolution, according to Morse, “we should be in a position to circumvent emerging diseases at fairly early stages.” Over the next two decades, a research program at the intersection of disease ecology and experimental virology coalesced around this goal.
Drawing on the assumptions of the “viral traffic” framework, international health authorities sought to implement a global surveillance system that could detect and rapidly contain novel pathogens. Epidemiologist Donald Henderson, who had led the W.H.O. smallpox eradication campaign in the 1970s, provided an initial vision of a global infrastructure for detecting the onset of emerging diseases. Henderson’s vision built on the field of “epidemic intelligence,” developed at the Center for Disease Control during the Cold War, which involved training field epidemiologists to track reports of outbreaks and quickly respond to contain them.
The goal was to extend the tools of epidemic intelligence — which were designed to detect outbreaks of existing diseases — to anticipate the emergence of novel ones. Such a system, Henderson and others argued, would be of use not only for monitoring emerging diseases, but also for the early detection of a bioterrorist attack. In the aftermath of the 2001 anthrax letters, this “dual use” approach to biological threats attracted increasing support from the national security establishment.
There were political as well as technical challenges to building a global system to monitor the emergence of infectious diseases. National governments were often hesitant to report outbreaks to international authorities. At the outset of the 2003 SARS epidemic, the Chinese government refused to allow outside experts to investigate, underlining the need for a system that would enable rapid response to the appearance of a novel and deadly infectious disease. Soon after, the specter of an avian influenza pandemic accelerated efforts to construct such a system, including the adoption of revised International Health Regulations that enabled the W.H.O. to declare a “public health emergency of international concern” based on the appearance of an as-yet-unknown pathogen. The goal was to push W.H.O. member states to rapidly report any such outbreaks and to allow international health authorities to investigate them.
A number of entrepreneurial scientists claimed to have developed tools that would enable the early detection of future viral disease outbreaks, making it possible to “stop the next pandemic before it starts.” Epidemiologist Larry Brilliant pitched a digital surveillance system that would crawl the internet and global news to detect signs of developing health threats. Primatologist Nathan Wolfe promoted a system of “viral forecasting” that involved collecting samples of bushmeat from local markets across sub-Saharan Africa. Zoologist Peter Daszak at EcoHealth Alliance conducted genetic analyses of samples taken from wildlife around the world with the aim of predicting the onset of emerging diseases.
The concern among health authorities that avian influenza would mutate to become more easily transmissible among humans drove an intensification of research on viral emergence. Disease ecologists thought that a pandemic strain would likely emerge at the duck-pig interface in East Asia and then be carried around the world by migratory waterfowl. To track potentially dangerous mutations of the virus, they regularly conducted genetic analyses of samples taken from migratory birds. But such molecular surveillance efforts posed a familiar question: How could scientists know which viral strains to look for? What were the signs that a virus was becoming more easily transmissible among humans?
Here a new set of scientific actors entered the picture: experimental virologists, who argued that it would be possible to simulate the natural process of viral evolution in the laboratory. The premise was that pushing H5N1 in the direction of human transmissibility would help molecular surveillance efforts by making it possible to identify genetic sequences linked to the ability to infect humans.
This goal was taken up as part of the U.S. government’s 2005 national pandemic preparedness plan. In an appendix to the plan, the National Institutes of Health pledged support for basic research in influenza virology, including projects to understand the “genetic changes that permit an influenza virus to suddenly acquire the ability to transmit between species.” This clause referred to a subfield of virology that would become known as “gain of function research,” in which researchers experimentally manipulated viruses in order to study characteristics such as virulence and transmissibility. Between 2001 and 2007, annual federal funding for basic research on influenza, managed by the National Institute of Allergy and Infectious Disease (NIAID), jumped from $15 million to $212 million.
Gain Of Function
When the H1N1 (swine flu) pandemic began in the spring of 2009, it seemed at first to be “the next pandemic” that health authorities had been anticipating. But H1N1 turned out to have relatively mild effects on human populations, and W.H.O. officials were accused of overreacting to the appearance of the new strain — of having pushed governments to invest huge sums in what turned out to be unnecessary mass vaccination campaigns. By this time, meanwhile, the dire threat of an avian influenza pandemic seemed to be waning. Perhaps H5N1 was unlikely, after all, to mutate to become easily transmissible among humans.
However, experimental virologists continued to make the case for its viability as a pandemic threat. In late 2011, when Dutch virologist Ron Fouchier announced that his laboratory had created a strain of H5N1 that could be passed via aerosol transmission among ferrets, he explained the rationale for creating what he described as “one of the most dangerous viruses you can make.” While there were respected scientists who thought “that H5N1 could never become airborne between mammals,” he said, “I wasn’t convinced. To prove these guys wrong, we needed to make a virus that is transmissible.”
Fouchier’s announcement sparked a public debate among life scientists and biosafety specialists over the risks and benefits of gain-of-function research on dangerous pathogens. According to advocates of the research, experimentally manipulating viruses to make them more virulent or transmissible would contribute to pandemic preparedness by enabling molecular surveillance efforts — such as sampling migratory birds for avian influenza — to recognize the emergence of dangerous pathogens in time to contain them. “In defining the mutations required for mammalian transmission,” as NIAID director Anthony Fauci and two co-authors wrote in a Washington Post op-ed, “public health officials are provided with genetic signatures that, like fingerprints, could help scientists more readily identify newly emergent, potentially harmful viruses, track their spread and detect threatening outbreaks.”
Critics, meanwhile, charged that the rationale for such research was tenuous at best. First, they argued, it took for granted a vast technical capacity for the molecular surveillance of potential animal hosts that was completely unrealistic given limited resources. Second, it assumed that what was created in the laboratory through gain-of-function research would in fact mimic what was likely to emerge in “nature.” But there was no basis for such an assumption. As one scientist memorably put it, “Would nature have come up with the dachshund?”
Critics also argued that a significant recent record of laboratory accidents resulting in the release of dangerous viruses and a woefully insufficient regulatory apparatus militated against providing government support for gain-of-function research. The key concern was that this type of research might spawn exactly what it was meant to prevent. As a group of scientists concerned with biosafety wrote in 2014, the lab-based creation of pathogens with pandemic potential “entails a unique risk that a laboratory accident could spark a pandemic killing millions.”
For these critics, the hypothetical benefit of assessing pandemic potential did not outweigh the catastrophic risk of unleashing an actual pandemic. Two catastrophic scenarios confronted one another with no means of technical resolution: a naturally emerging virus whose onset might be anticipated and even prevented through the results of viral transmission research, versus the accidental release of a pandemic virus as a result of this very research.
In this uncertain terrain, government funding agencies struggled to find an agreed-upon method of risk assessment that could guide regulatory decisions. Meanwhile, despite an official moratorium on federal support for gain-of-function research from 2014 to 2017, such experimentation continued and extended to new areas.
The Route To Wuhan
The two major strands of research on viral emergence — disease ecologists studying wildlife in the field and experimental virologists manipulating pathogens in the laboratory — converged in the investigation of bat coronaviruses found in caves in southern China. In June 2014, NIH funded a proposal for research on “understanding the risk of bat coronavirus emergence,” led by Peter Daszak of EcoHealth Alliance in collaboration with the Wuhan Institute of Virology. The proposed research would address questions on “the origin, diversity, capacity to cause illness and risk of spillover” of bat coronaviruses, and involved “conducting laboratory experiments to analyze and predict which newly discovered viruses pose the greatest threat to human health.”
Such “emergence potential” — that is, the potential for “interspecies transmission” of novel coronaviruses — would be tested “using reverse genetics, pseudovirus and receptor binding assays, and virus infection experiments across a range of cell cultures from different species and humanized mice,” as the proposal put it. Thus, the project sought to address the question that Stephen Morse had posed three decades earlier: How to assess “the likelihood that a given animal virus will emerge as a pathogen”?
Seven years later, in mid-2021, there were two ways to understand the retrospective significance of this research program — as either the prescient forecast or the dangerous progenitor of the COVID-19 pandemic. In this sense, we can understand our current situation of diagnostic uncertainty as a question of which route of viral traffic to follow: zoonotic spillover, as exemplified by the movement of SARS in the early 2000s from bats to civet cats to humans via the trade in wildlife; or a new potential route, from the bat caves of southern China to a virology laboratory in Wuhan, as part of a cosmopolitan project in the life sciences—initially proposed in 1989—to investigate the pandemic potential of emerging viruses. The stakes of this assessment are high, not only for determining the sites of failure that led to the present catastrophe, but also in targeting interventions designed to forestall “the next one.”