A Self-Fulfilling Prophecy Pandemic The 1977 "Russian Flu"

Abstract

Surprisingly, the 1977 "Russian flu" H1N1 pandemic influenza virus was genetically indistinguishable from strains that had circulated decades earlier but had gone extinct in 1957. This essay puts forward the most plausible chronology to explain the reemergence of the 1977 H1N1 pandemic virus: (1) in January–February 1976, a self-limited small outbreak of a swine H1N1 influenza virus occurred among Army personnel at Fort Dix, New Jersey; (2) in March 1976, the US launched a nationwide H1N1 swine influenza vaccine program; (3) other countries then also launched their own H1N1 R&D efforts; (4) a new H1N1 outbreak, genetically unrelated to the Fort Dix swine virus but indistinguishable from previously extinct H1N1 viruses, was detected early in 1977 in China; (5) the leading Chinese influenza virologist later disclosed that the Chinese military had conducted large H1N1 vaccine R&D studies in 1976. It is likely that the resurrected H1N1 influenza viruses were laboratory-stored strains that were unfrozen and studied as part of the emergency response to a perceived epidemic threat, and that accidentally escaped. The fear of a possible H1N1 pandemic was the critical factor that gave rise to the actual H1N1 pandemic, resulting in an avoidable "self-fulfilling prophecy pandemic."

The study of epidemic origins is a relatively young scientific pursuit. As recently as 1989, Nobel Laureate Joshua Lederberg (1989), self-critically commenting on scientists' seeming incuriosity about how epidemics begin, wryly observed that "the historiography of epidemic disease is one of the last refuges of the concept of special creationism." Lederberg coined the term "emerging infectious diseases" for this new field of study, emphasizing that these diseases must originate—or emerge—from somewhere. The hope was that by carefully studying the origins of past pandemics, future pandemics could be prevented. The 1977 H1N1 influenza pandemic is a particularly instructive case.

The 1977 "Russian Flu": An Odd Pandemic

As pandemics go, the global spread of H1N1 influenza in 1977 wasn't especially fearsome (CDC 2010; Zimmer and Burke 2009). The "Spanish flu" that circulated between 1918 and 1919 is the most well-known, and for good reason (Taubenberger and Morens 2006). That particular swine flu virus is estimated to have infected roughly 500 million people, which made up one-third of the world's population at the time. At least 50 million people died worldwide, with 675,000 deaths in the US alone. Other major pandemics of the last century include the 1957 "Asian flu" and the 1968 "Hong Kong flu," each of which killed roughly a million people. By comparison, the 1977 "Russian flu" virus produced a milder illness in most people (Gregg, Hinman, and Craven 1978). Over the ensuing years, after its emergence in 1977, the virus went on to infect a significant portion of the worldwide population, killing roughly 700,000 people. But this particular pandemic isn't important only because of its impact on disease and death: it also reveals uncomfortable truths about the complex relationship between viruses and humans. The 1977 pandemic is a cautionary tale, one that—hopefully—we can learn from.

In 1978, Zdanov and colleagues described a new H1N1 influenza virus epidemic that had begun in Russia late in the previous year (Zakstelskaja et al. 1978). The epidemic was first reported to the World Health Organization (WHO) by Russia on December 7, 1977. The first case was that of a 22-year-old man in Moscow—the on-paper official "patient zero" of the 1977 H1N1 pandemic. Shortly thereafter came a report of an outbreak in a group of young fishing apprentices in the eastern village of Nakhodka. By early December, roughly a month after the Nakhodka outbreak, all 50 major cities in the USSR's influenza surveillance system were reporting similar cases (Gregg, Hinman, and Craven 1978).

The virus became known as the "Russian flu," but that was a misnomer, we now know, as it had actually first appeared earlier, during the spring of 1977, in the port city of Tientsin, China. There, it emerged in teenagers and military populations and then spread to preschoolers before infecting the general population. Chinese public health laboratories isolated the virus and identified it as H1N1, later reporting it to WHO in January 1978 (Kung et al. 1978). Soon it had spread to the United Kingdom, and then it popped up outside of Paris, next in Istanbul. It was first detected in the US in January 1978, in Cheyenne, Wyoming, where it ignited an outbreak at a local high school that only involved students (Kendal et al. 1979). The virus also flourished in other schools and military bases throughout the country. That winter, 1977–78, a global pandemic was well underway.

This new H1N1 influenza virus displayed several unusual features. First was its relatively low mortality impact. US researchers have estimated that in the decade after the virus' introduction, the number of influenza-associated deaths during predominantly 1977 H1N1 years was substantially lower than in predominantly H3N2 years and even predominantly influenza B years. (Three predominant H1N1 years average = 11,892 deaths; three predominant B years average = 15,479 deaths; four predominant H3N2 years average = 36,848 deaths; Thompson et al. 2003). The second peculiar feature of the 1977 Russian flu was that older people were immune to it, and it mostly affected people under the age of 26, while traditionally flus have their greatest impact on the elderly (Gregg, Hinman, and Craven 1978). And third, the historical pattern prior to 1977 was that when a new influenza subtype emerged, it displaced the currently circulating influenza subtype (Monto and Fukuda 2020). But the new H1N1 virus continued to circulate alongside the more longstanding H3N2 strain. It was an odd pandemic in many ways.

Soon the story took an even more puzzling twist. Peter Palese, a microbiologist at the Icahn School of Medicine at Mount Sinai in New York, was well on his way to becoming a leading expert in RNA viruses when he and his colleagues in 1978 published a breakthrough paper in Nature. The paper reported the use of a then-new genetic analysis technology called oligonucleotide mapping to study the 1977 Russian and Asian H1N1 strains (Nakajima, Desselberger, and Palese 1978). Compared to modern sequencing technology, the basic technique was cumbersome (Brownlee and Sanger 1969). Rather than directly amplifying and sequencing the influenza virus RNA, as is done now, the researchers first digested—chopped—the viral RNA into pieces with an enzyme called T1 RNase, then used electrophoresis to drive the RNA cleavage fragments apart from each other in two dimensions, generating a flat, spotted pattern. By comparing the RNA spot movement patterns (known as "maps") of two viruses, researchers could determine the viruses' genetic similarity or dissimilarity based on the congruence of their RNA maps. Although the method was complicated, it was also quite sensitive. As Palese told me more than 40 years after his experiment, as little as a 5% percent difference in the RNA sequences of two viruses would result in completely different oligonucleotide maps (personal interview, Nov. 23, 2021). This experiment not only ushered in a new technology for investigating the pandemic, but it launched a profound shift in the way we see and understand viruses themselves. Palese and his colleagues titled their paper "Recent Human Influenza A (H1N1) Viruses Are Closely Related Genetically to Strains Isolated in 1950."

Escape from Plato's Cave: The Birth of Molecular Epidemiology

Scientists knew at the time of Palese's study that the immune system "sees" a virus through antibodies, which are proteins that have complementary shapes to viral molecular structures. But antibodies offer only indirect measurements, and there were no scientific methods to examine the details of the actual virus structure. It was as if virologists were trapped in Plato's cave, where they could only view the immunological shadows on the wall and not the reality of the virus itself. Their ability to follow the evolution of virus structures was limited to measuring changes in the hazy patterns of antibody shadows.

These technological limitations also blurred our ability to compare viruses. When two viruses are similar to each other, antibodies will cross-react to their proteins. To measure the degree of similarity between viruses, scientists would make dilutions of antibodies to see at which end point they still reacted with a virus. If the antibodies could be diluted multiple times and still react, the viruses were more similar. If the cross-reactions stopped after a few dilutions, the viruses were more different. The result was always an indirect measure. But in the late 1970s, it became possible to directly compare viral RNA patterns and visualize the genetic characteristics of a virus.

Up until this scientific moment, it had been difficult to figure out where viral strains came from. But suddenly, you could make sense of things that otherwise would have been foggy using antibody patterns. And yet it was hard for Palese and his team to make any sense of what they saw during that pivotal experiment in 1978. Curiously, they found that RNA maps of the 1977 strain were essentially identical to an extinct H1N1 strain that had circulated between 1946 and 1957, then disappeared. According to Palese, the two H1N1 strains, one that had recently caused a pandemic and one that had been extinct since 1957, were "basically indistinguishable."

This finding defied a natural explanation. The number of influenza virus particles on this planet is unthinkably large, on the order of millions of billions, and all these viruses are relentlessly mutating and reassorting their RNA. Like snowflakes, every newly replicated influenza virus is unique. Each virus records its genetic information on small chromosome-like segments of RNA and nucleoprotein, just as humans do, except viruses don't get half of their chromosomes from each of two parents. Instead, they have eight distinct RNA segments, and when a virus infects a cell, all eight segments are replicated and packaged into new progeny virus particles. Occasionally, when two influenza viruses infect the same cell, they can swap any one of these eight segments, or none at all. Segment swapping, or "reassortment" can lead to new hybrid progeny influenza viruses. And with or without reassortment, mutations continue to accumulate with every successive replication. "We already knew that there was an increase in mutations over time," Palese says of his reaction to the findings, "so suddenly when [a virus] goes back to 1950, it was very, very strange" (personal interview, Nov. 23, 2021).

In their Nature paper, Palese and colleagues concluded that, "If we accept the premise that influenza A viruses are subject to repeated mutational events it is extremely difficult to explain why the oligonucleotide maps of strains isolated in 1950 and those of the recent Russian viruses are so strikingly similar" (Nakajima, Desselberger, and Palese 1978, 339). Antigenic drift, the genetic variation that arises from mutations, would have occurred had the virus been continually transmitted by humans, they argued. And while reverse mutations, whereby a virus regains its previous genetic properties, would not have been impossible, they certainly weren't probable. "It seems much more likely," the authors concluded, "that the genetic information in the Russian viruses has been preserved over the last 25–27 years by some unusual mechanism" (339). Perhaps, they posited, the virus had been contained in some kind of animal where it didn't undergo the usual genetic changes, or frozen somewhere in nature. The paper leaves the questions of origin open.

The same year of Palese's experiment, Chi-Ming Chu, the leading influenza virologist in China who would go on to establish the field of molecular virology and become a lauded expert in influenza research, was also studying the 1977 H1N1 virus in Beijing (Yan et al. 2017). In a 1978 paper published in The Bulletin of the World Health Organization, Chu and his team noted the same similarities between the 1950 and 1977 viruses: "The reappearance of H1N1 virus after its apparent disappearance for 20 years, resulting in widespread epidemics, is an unprecedented event in the history of influenza" (Kung et al. 1978, 913). Chu and colleagues set out to understand the virus's origin, confirming that it mostly attacked people younger than 20 years old. He speculated, like Palese, that perhaps the latent virus had been somehow reactivated after lying dormant. In his conclusion, Chu cited the then-prevailing "recycling" theory of influenza viruses—the notion that influenza subtypes, like H1N1, tend to reemerge in cycles every 70 to 80 years. He stated that the appearance of an epidemic H1N1 virus roughly 60 years after the notorious 1918 H1N1 Spanish flu was "more than accidental." Perhaps, in other words, it was right on time. But Chu didn't offer a definitive answer to explain the similarity between the two H1N1s. No one could: there simply wasn't enough information at the time.

Clever Viruses

The 1978 studies of H1N1 strains by Palese and Chu raised more questions than they answered. If viruses are nothing but short segments of genetic code packed into tiny particles, how do they periodically emerge to cause pandemics? It may be useful to conceptualize viruses as possessing a type of primitive intelligence. Joshua Lederberg, the Nobel-winning molecular biologist who discovered bacterial transduction, used to say that when it comes to our relationship with viruses, "It's our wits vs. their genes" (Lederberg 2000, 290). Napoleon's surgeon general spoke of the "genius" of epidemics, in both senses of the French word, génie (Debre 2000). Viruses are genies in a bottle: they're magical. But they are also the kind of genius associated with intelligence, in this case a proto-intelligence dedicated to replication and infection. For millions of years, influenza viruses have been perfecting the art of replicating their 10,000 nucleotides and finding new ways to jump from one species to another. Every one of them is constantly mutating and reassorting, and in the process adapting to their environment. The strategies they use are not unique to viruses: humans have developed computer programs that employ these same evolutionary problem-solving methods as a type of artificial intelligence known as "genetic algorithms." The viruses use strings of nucleotides, whereas humans use strings of computer code (Burke et al. 1998; DeJong, Spears, and Gordon 1993). After several hundred million years, viruses have become very good at this game. The viruses are continuously testing out new genetic sequences, relentlessly probing the possibilities. And sometimes they hit the jackpot.

In return, we humans are playing our side of the game. We are constantly probing viruses—testing, learning, tracking, experimenting. And sometimes we get duped.

Flashback: Fort Dix, 1976

To understand the origins of the 1977 H1N1 pandemic, we have to go back in time roughly one year, to January 1976 (CDC 1976). Specifically, to the Fort Dix Army post in central New Jersey, where roughly 6,000 new recruits arrived shortly after the Christmas and New Year's holidays. Together they spent three intense days at the reception center, taking exams and completing their registration. Basic training resumed on January 5 and would last for seven weeks. That day the wind-chill factor dropped to –43° Fahrenheit, and shortly thereafter soldiers started to report respiratory symptoms: coughing, shortness of breath, and lingering chest pain. On January 19, a recruit was hospitalized with acute respiratory disease. More hospitalizations followed, and healthcare workers began to collect throat swabs. A week later, 19 specimens had been delivered to the state laboratory, which identified seven A/Victoria-like viruses (the main source of human influenza since 1968), but also three unknown viruses (CDC 1976; WHO 1976).

Then, on February 4, 19-year-old Army recruit David Lewis went out with his platoon on a 50-mile overnight training hike. After 13 miles trudging through the cold, Lewis collapsed, and he died later that evening. An autopsy identified pneumonia as the cause of death, and postmortem tracheal swabs inoculated into embryonated chick eggs grew an unknown virus, which was found to be swine influenza.

Figure 1.

Private David Lewis, high-school yearbook photo

Private Lewis was born in Buffalo, New York, on October 18, 1956, the middle of three children. His mother Roberta Jane was a registered nurse who worked in coronary care, and his father Rev. Harry W. Lewis was a Baptist minister and Korean War veteran. Lewis attended the Mount Everett Regional High School in southeastern Massachusetts, graduating in 1974 (see Figure 1). He played freshman football, was a member of the chess club, and tutored younger elementary students. After high school he attended and obtained a degree from the Word of Life Bible Institute in central New York, an organization that prepares students for the ministry. Exactly why he joined the Army in 1976 is unknown. The Vietnam War had recently ended, and the military had converted from a draft to an all-volunteer force. High-school friend Tom Berkel was surprised when David voluntarily enlisted. Tom described him as "smart as a whip" and thought he would be "college bound" (Berkel, personal interview, 2021; Berkshire Eagle 1976; Mt. Everett Regional HS 1974; Wilkes-Barre Times Leader 1976). Lewis's death—the only death from swine flu—stoked fears of a deadly pandemic and helped to set off a global immunization scramble that did not end well.

Swine flu is usually confined to pigs, where it causes influenza illness with familiar symptoms: runny nose, congestion, cough, and respiratory problems. There are many influenza viruses adapted to particular animal species in nature—duck flu, horse flu, bat flu—and they have evolved to target those specific animals. Occasionally an animal virus will jump from animals into humans. How is this possible? The main contributing factor is proximity. For a virus to jump species, there must be intimate contact plus a mutation or recombination that makes the virus transmissible. On average, an RNA virus makes about one error in copying the 10,000 nucleotides in its genome each time it replicates (Drake and Holland 1999). So if one virus isn't the exact right fit for jumping to another species, then perhaps the next one will be, or the next one, or the one after that. These types of jumps don't occur easily, but when they do, the human population is virgin viral soil.

Virologists in 1976 were fearful that if H1N1 swine flu were to suddenly return to humans, people under the age of 19 would have no antibodies against infection, because they hadn't been exposed to any previous H1N1 influenzas in their lifetimes. Most alarming, of course, was the threat of a repeat pandemic, especially the specter of 1918. Until October 2021, when deaths from COVID-19 surpassed 700,000, the 1918 Spanish flu was still the deadliest pandemic in American history. The same H1N1 strains continued to circulate after the 1918 pandemic, until genetic reassortment generated the H2N2 pandemic of 1957 (Zimmer and Burke 2009). The young Fort Dix population was especially susceptible to H1N1: of the 19,000 people at the infantry training facility, roughly a third of them were recruits in their teens or early 20s. They were too young to have been previously exposed to H1N1 prior to the emergence of the H2N2 virus in 1957, and for that reason they were vulnerable in 1976. Another concern was that Fort Dix was within close range of Philadelphia and New York, both densely populated regions where the new virus could flourish. Any responsible epidemiologist had to take this threat seriously.

Colonel Franklin Top is a Yale medical graduate and infectious disease expert who served as Chief of the Department of Virus Diseases at Walter Reed Army Institute of Research during the Fort Dix outbreak. After his distinguished military career, Top joined the biotech firm Medimmune, Inc., and went on to lead development of Medimmune's successful FluMist nasal influenza vaccine, HPV vaccine, and RSV monoclonal antibodies. I (DSB) spoke to him some 45 years later about receiving a call from the CDC on February 14, 1976, and flying to Atlanta for a meeting that same night (personal interview, July 19–20, 2021).

"We recognized it was going to be a real problem," Top remembers. He and his boss, Philip K. Russell, knew that the term "swine flu" would be troublesome because of the 1918 pandemic, and they assumed "that it would hit the papers, that there would be enormous interest in what was going on by the government … [and] we realized we were going to be under a lot of pressure to deliver." That same day, February 14, the Morbidity and Mortality Weekly Report described "increases of febrile disease at Fort Dix" during the last two weeks of January. Of David Lewis's death, the report stated that "Postmortem findings were consistent with pulmonary viral infection." Regarding the infections then on record, the report stated that "None of the cases had known contact with swine" (CDC 1976).

At the CDC, a group of experts was forced to grapple with how to handle the emerging cases at Fort Dix. Was a pandemic on the horizon, or would the virus infect a few people and then retreat? Everyone agreed on one frustrating point: there was simply no way of knowing for sure. CDC Director David Sencer held a press conference on February 19, and as Top had predicted, the media took the story and ran with it. "The possibility was raised today that the virus that caused the greatest world epidemic of influenza in modern history—the pandemic of 1918–19—may have returned," wrote Harold M. Schmeck Jr. in the New York Times. But Top couldn't let fears or speculation affect his work. He spent seven days a week in the lab to deliver reliable data that would help the CDC make these tough decisions. He and his team kept identifying A/Victoria (H3N2), the more prevalent influenza virus at the time, and had to tease out the A/New Jersey H1N1 infections from this background of other influenzas in order to get a better grasp of the novel swine flu's reach.

When Sencer reconvened the same group of scientists and government officials on March 10, there was still conflict about how to proceed, according to Richard Neustadt and Harvey Fineberg in The Swine Flu Affair: Decision-Making on a Slippery Disease (1978). If they overreacted to A/New Jersey by producing an unnecessary vaccine, the CDC would be to blame. But facing a possible pandemic without a vaccine was perhaps worse. A/Victoria vaccines were already being manufactured at that time, as the CDC had recommended the vaccine for people over 65 with certain chronic conditions. Now they had to make a quick decision about producing a vaccine against A/New Jersey. They would need new supplies and extensive testing in field trials. Time was not on their side.

Sencer became convinced that the government should proceed with a nationwide vaccination campaign. On March 24, alongside Albert Sabin and Jonas Salk, two respected scientists famous for developing polio vaccines, President Gerald Ford announced his decision to appropriate funds to produce a swine flu vaccine and "inoculate every man, woman, and child in the United States." Ford acknowledged that "no one knows exactly how serious this threat could be," but also claimed that the country couldn't "afford to take a chance with the health of our nation."

The research being done by Top and colleagues, however, showed a rapidly dwindling threat. Overall, 230 soldiers at Fort Dix had been infected with the A/New Jersey virus; 13 were hospitalized with respiratory disease, and just one, Private Lewis, had died. In a 2006 paper in Emerging Infectious Diseases, Top and his team from Walter Reed wrote that the "numbers of isolation and serologic specimens tested and the percentages positive for A/Victoria were consistent with an outbreak that began quickly in January and declined in late February to early March. No influenza cases were identified after March 19; influenza A/New Jersey was never isolated outside Fort Dix" (Gaydos et al. 2006, 27). In other words, the A/New Jersey swine flu virus simply disappeared.

Epistemic Hubris

Although the new swine flu outbreak at Fort Dix appeared to have self-extinguished, the US government's vaccine machine was already well underway. Field trials began on April 21. "Frankly, I was sort of surprised," Top says, "because I thought we had shown pretty clearly that [the virus] didn't go anywhere but Fort Dix. Somebody kissed their pig goodbye, got into Dix, and passed it on for about three weeks, and then it disappeared." He and Russell had concluded that a vaccine likely wasn't needed because they couldn't find the virus anywhere. Top did agree that field trials and clinical studies were appropriate, because the virus may have popped up somewhere else. "The wisest thing to do was to get a vaccine program started," he says, "but I didn't think it made an awful lot of sense to go ahead and immunize after we hadn't seen [the virus] again. … I was in fact a little leery that we ultimately would be blamed for immunizing a good part of the population with a vaccine for which there may not have been a virus."

Peter Palese, who conducted the RNA map study on the 1977 influenza virus, was also involved in the Fort Dix outbreak because he worked for Edwin Kilbourne (2006), a medical researcher and flu expert who was concerned about the virus spreading and strongly in favor of vaccination. Palese pointed out similarities between the 1976 swine flu virus and several other swine H1N1s that had been isolated at Oscar Mayer factories but never took hold in human populations. He didn't think the Fort Dix virus posed a threat and recommended caution: "You don't vaccinate for a non-disease." But Kilbourne argued that you couldn't get the government to actually produce the vaccine if you weren't going to distribute it.

In the end, the government stuck to its plan, and on October 1 launched a mass immunization campaign. Within weeks, new problems arose. Three people over 70 who had cardiac conditions died shortly after receiving the vaccine at a Pittsburgh clinic. Despite President Ford and his family receiving the vaccine on live television just a few days later, doubt had started to set in. Still, between October 1 and December 16, more than 40 million Americans had been vaccinated against A/New Jersey swine flu. Of those 40 million civilian vaccinations, there arose 532 cases of Guillain-Barre syndrome, a rare side effect in which the immune system attacks the nerves; 32 people died from it (Langmuir et al. 1984). The program was officially suspended on December 16, 1976. Other than a single swine flu case in Missouri that was not traceable to pigs, no cases of the A/New Jersey were ever identified outside of Fort Dix.

Epidemics are unique, highly chance-driven events. They almost uniformly defy prediction. You can never be certain whether a particular virus will take hold, but in 1976, the decision-makers acted with far too much certainty.

So what happened to the novel swine flu virus from Fort Dix? One of my (DSB) outstanding graduate students at Johns Hopkins, Justin Lessler, now a professor in the Department of Epidemiology at the University of North Carolina, dug deep into the military epidemiological data on the location and timing of swine flu infections at Fort Dix and simulated the likely transmission patterns and the intrinsic transmissibility of the virus (Lessler et al. 2007). Early on, the young Fort Dix trainees were grouped into 50-member platoons and then organized into companies, each with four platoons. Within platoons, members had close contact, but they had less contact with recruits in their company, and even less with those in other companies. All of this created barriers to transmission. In a 2007 paper we published in the Journal of the Royal Society Interface, we estimated the basic reproduction number of A/New Jersey to be 1.0 to 1.2, just barely over the critical transmissibility threshold for sustained transmission, even in the lush virological breeding grounds of a multitude of susceptible young men, stressed and crowded together, in the dead of winter (Lessler et al. 2007). Lessler also estimated that the virus had been serially transmitted person to person approximately six times before it died out. We realized that this had been one of "nature's false starts." The swine virus likely ran its course: it simply wasn't fit enough to be transmitted among humans outside the exceptionally virus-friendly environment at Fort Dix during that cold winter of 1976. It is possible—even likely—that the virus might have infected a few people in the New Jersey community around the base, but no cases were ever detected in community screening studies done by the CDC. If they occurred at all, any infections outside the base led to short, self-terminating chains of transmission. And if the virus did adapt to humans during its short cross-species exploratory foray on the military base, it was not enough adaptation to permit sustained transmission in civilian populations. It was a false start, to be sure, but one that had far-reaching implications, well beyond the borders of Fort Dix.

Connecting the Dots

My (DSB) own professional experience with the risks of human error and laboratory infections may be the reason why, as far as I know, I was the first person to connect the dots between the 1976 Fort Dix swine flu outbreak and the subsequent 1977 H1N1 pandemic. I wrote an op-ed for Asian Wall Street Journal in January 2004 in which I discussed the first SARS coronavirus and the larger threat of cross-species transmission (Burke 2004). Emerging diseases seemed to be popping up everywhere at that time, with a new bird flu rampant among poultry in Southeast Asia and the novel coronavirus, SARS-CoV-1, showing sustained human transmission. Here is what I wrote about my concern that research on these viruses might contribute to the kinds of accidents and mishaps I had observed throughout my career:

Unintentional release of extinct human-adapted viruses—"bio bungling"—arguably poses as serious a threat to global health as bioterrorism or a natural outbreak. Again, an example from influenza is instructive: Genetic sequencing of the global pandemic 1977 H1N1 influenza virus has shown it to be identical to an H1N1 strain that became extinct outside laboratories in the 1950's. The most plausible scenario is that the 1977 virus was one stored for decades in a laboratory freezer and thawed for experimental study during the 1976 swine influenza scare: a "self-fulfilling prophecy" epidemic. (emphasis added)

By then it was clear to me that scientific experimentation with extinct viruses—no matter how carefully done—could pose a greater threat than emergence from nature. An extinct epidemic virus is one that has already proven its ability to be efficiently transmitted from person to person and is an immediate threat to causing a renewed epidemic: a hair-trigger microbial bomb waiting to go off. By comparison, a newly emerged virus must first adapt to humans before it can spread efficiently. Risky, but less so.

In December 2004, roughly a year after my op-ed was published, Peter Palese wrote an article in Nature Medicine Supplement about influenza genetics and the risks posed by research. In it, he revisited his earlier attempts to pin down the origin of the 1977 H1N1 virus. "Although there is no hard evidence available," he wrote, "the introduction of this 1977 H1N1 virus is now thought to be the result of vaccine trials in the Far East involving the challenge of several thousand military recruits with live H1N1 virus," citing his personal communication with Chi-Ming Chu (S83). Palese had personally related to me the possibility that the 1977 H1N1 had been accidentally released during research studies in China, but this was the first time he declared it in print. Interestingly, in his 1978 Nature paper, Palese had not mentioned the intentional opening of a long-frozen laboratory vial, or escape during vaccine trials involving intentional human challenges, among the possible modes of resurrection and reemergence of the 1950 virus.

Why would anyone intentionally work with an extinct virus? The most logical motivation would be to prepare for and prevent the possible emergence of a virus closely related to the extinct virus—that is, by creating in advance a vaccine against the anticipated virus. Although Palese doesn't specify a country, only referring to the location of the trials as having taken place "in the Far East," the fact that the information was told to him by Chu, a Chinese scientist, coupled with the fact that the resurrected H1N1 virus was first recognized in China, make it very likely that Dr. Chu was referring to vaccine studies that must have taken place in China. The years 1976 and 1977 were a tumultuous time in China, as Mao Tse-tung had just died and Deng Xiaoping was repudiating the Cultural Revolution, so records of exactly what was done may be difficult to find. Nonetheless, we can speculate about what might have happened.

The first possibility is that live 1950 virus was grown in large quantities in chicken eggs and then used in challenge studies to inoculate human subjects who had been vaccinated against H1N1 viruses, as a way to directly test the protective efficacy of new vaccines. Even if the human subjects had been held in isolation during such trials, it is possible that the 1950 H1N1 virus could have escaped into the general population. The sheer size of the trials as claimed by Chu—involving thousands of persons—make an inadvertent virus escape plausible.

A second possibility is that the 1950 virus was grown in quantity in chicken eggs and used as the starting material for a new vaccine, and that faulty preparation led to accidental release of hot live virus. For example, if an influenza vaccine was being developed as a traditional whole inactivated "killed virus" vaccine, it is possible that not every virus particle in the research vaccine preparation was successfully inactivated. Exactly this problem—residual infectious virus in a "killed virus" vaccine preparation—had happened in the US in 1955 during the initial rollout of the inactivated Salk polio vaccine, when a lot of 120,000 incompletely inactivated doses from Cutter Laboratories was administered to children, resulting in 56 cases of paralytic polio and five deaths (Offit 2005). Even more tragically, the Cutter vaccine live polio virus spread from vaccine-infected recipients to their families, causing another 113 cases of paralysis and five deaths.

A third possibility is that attempts had been made to create a live attenuated replication-competent vaccine from the resurrected 1950 virus, and that this 1950 strain-derived live vaccine virus could have back-mutated to full virulence and begun to spread.

Experimental Human Infections

Fifty years ago, experimental human inoculations were commonly done to study infectious diseases, especially self-limiting diseases that posed minimal risk, or bacterial infections that were fully treatable with antibiotics. In 1973, when I (DSB) joined the US Army Medical Research Institute of Infectious Diseases (USAMRIID), the biodefense medical research program known as Operation White Coat that was carried out at the base for nearly 20 years had just been shut down (Pittman et al. 2005). The draft had ended, and the Seventh-Day Adventist Church conscientious objectors who made up most of the White Coat volunteers were no longer available. Roughly 2,300 young men had volunteered for one or more studies in which they were infected with pathogens that had potential for biological attacks. There were no resulting deaths or long-term consequences. To my knowledge, no influenza studies were done through that particular program, which was focused on biodefense vaccines, but the program as a whole illustrates that human inoculation studies were commonly conducted in the US. I have no direct information about similar human inoculation studies in military populations in China or Russia, but it is likely that they occurred, possibly with less attention to informed consent than in the US.

The Common Cold Unit of the British Medical Research Council, whose primary mission was researching the common cold, conducted several important human challenge studies with live influenza virus during the 1960s and 1970s. In 1971, five years before the Fort Dix outbreak, they reported results of a study of human inoculations with viruses originally isolated from swine, showing that 11 of 34 inoculated human volunteers became infected with swine viruses, the majority of whom shed virus and were therefore potentially contagious (Beare et al. 1971). This seminal paper presciently concluded that "large outbreaks of human infections may be caused by major variants originating in animals." Immediately after the Fort Dix outbreak, in July 1976, this same Common Cold Unit published a report of human challenge studies with the already-extinct A/New Jersey virus, among other extinct flu viruses, and found that six of six human volunteers became infected (Beare and Craig 1976). These studies were carefully done on limited numbers of volunteers housed in isolation wards, and fortunately no viruses escaped into the general population.

In retrospect, these experimental human infections had skated awfully close to the edge of catastrophe. The viruses in these studies had all ceased circulating among humans many years earlier. The combined effect of a yearly birth rate of about 2% of the population, coupled with a rapidly waning immunity to influenza among the population previously exposed to these strains, had generated a growing pool of susceptible humans. Had any one of these resurrected viruses escaped the research ward, the population would have been vulnerable to the rapid spread of these already human-adapted viruses.

Currently in the US, human live influenza virus challenge studies are being done at four NIH-sponsored units at the University of Maryland School of Medicine, Saint Louis University, Duke University, and Cincinnati Children's Hospital (NIH 2019). None of these sites are experimenting with extinct pandemic strains.

A Self-Fulfilling Prophecy Pandemic

If we accept Chu's statement at face value, that the 1977 flu strain was derived from human vaccine trials in China conducted to prepare for a possible future H1N1 epidemic, we still must understand the logic. Why this particular virus? Why this exact point in time?

"I'm convinced that it was not done deliberately," Palese told me when I asked him to elaborate on Chu's account. According to Palese, Chu informed him of the H1N1 vaccine trials at a meeting in Russia sometime after 1978. From what Palese recalls, it was a distinctly military operation, and he has no reason to believe Chu was personally involved. "I think [the military] just didn't even conceive of the idea that the vaccine may not be as good," he said. "I'm pretty sure that this was not done in any malicious way." Palese's conversation with Chu is only testimony—not evidence or proof.

Decades later, in 2010, phylogeneticist Joel Wertheim at University of California, San Diego, analyzed sequences of the 1977 H1N1 virus to more precisely determine its moment of origin. Using a "molecular clock" method, which looks at the rate of mutations to determine when a viral RNA sequence emerged from a common ancestor, Wertheim found that "the 1977 reemergent lineage was circulating for approximately one year before detection," placing its emergence in roughly May of 1976 (e11184).

The emergence of a resurrected 1950 H1N1 influenza strain in China just months after the self-limited Fort Dix H1N1 swine influenza outbreak is a highly improbable coincidence, strongly suggestive that it wasn't by chance. Indeed, the estimated emergence of H1N1 in China in May 1976 almost exactly matches the moment that vaccine studies with the Fort Dix swine flu virus were initiated by the US government—in late April 1976.

Why would the Chinese government decide, in the spring of 1976, to develop a vaccine against an extinct H1N1 influenza virus? No doubt the Chinese military were closely following the dramatic events in the US that had begun on a military base and promptly led to a crash national vaccination program. Chinese health officials likely took the Fort Dix outbreak, though self-limited, as a harbinger of the reemergence of H1N1 viruses. Edwin Kilbourne, Robert Webster, and other leading American influenza virologists of the day subscribed to the theory that there were a limited number of influenza virus types in nature, and that these virus types cyclically reemerged and dominated every 60 to 70 years (Francis 1952). In 1976, almost 60 years had passed since the great H1N1 influenza of 1918, so it was thought that H1 was again "due" to reemerge. Chu also subscribed to this view in his 1978 paper, along with his American contemporaries.

If there were any remaining questions about the similarities between the 1950 and 1977 H1N1 viruses, they were put to rest in 2015. Rozo and Gronvall (2015) used modern genetic technology to reveal that the 1977 strain was 98.4% identical to H1N1 influenza virus strains isolated between 1948 and 1951, with only four differences among 566 amino acids. This match is so mathematically close that the inconsistencies are likely due to a sampling error alone.

All evidence points to the 1977 H1N1 Russian flu as a "self-fulfilling prophecy pandemic." We, the entire human species, were so worried in 1976 that H1N1 influenza would reemerge that we ourselves made it happen. The virus—a clever adversary—made a head fake at Fort Dix, and we fell for it. Our urgent efforts to understand and thwart the virus, however well intentioned, directly led to the resurrection and reemergence of the virus and a global pandemic that affected millions, if not billions. In the virological jiu-jitsu of our wits versus their genes, the viral genes won that round.

For those who seek to assign blame, there is plenty of it to go around. All of humanity, rather than any individual scientist or country, has some measure of culpability. The Americans overreacted by launching a nationwide vaccination program against a virus that had already disappeared, raising global alarm. The British took risks with a variety of extinct but already human-adapted influenza viruses by inoculating them into volunteers, and by so doing, normalized human experimentation with resurrected viruses. The Chinese reportedly conducted vaccine trials that involved challenging human subjects with resurrected influenza. And it seems likely that scientists in other countries may have undertaken other risky studies. Because the virus was first detected in China, and because Chu told Palase that it was 'the result of vaccine trials in the Far East," an origin in China is most likely (Palese 2004). But it could have happened anywhere.

The Chernobyl No One Noticed

On April 26, 1986, the Number 4 reactor at the Chernobyl Nuclear Power Plant in northern Ukraine failed during an equipment test, resulting in a core meltdown and several explosions. Between 9,000 and 16,000 fatalities have been directly or indirectly linked to the disaster, which left 1,000 square miles uninhabitable (Beresford et al. 2016). The Chernobyl disaster was one of the most iconic failures of technology ever known. In stark comparison, no one at the time realized that the 1977 pandemic, which killed 700,000 people and encircled the globe, had been a catastrophic technology failure, one of even greater impact than Chernobyl.

The thought was never entertained by most scientists that the pandemic could have been the result of human error. Of course, the scientific tools available at the time were insufficiently precise to establish the near perfect identity between the 1977 strain and the extinct 1950 strain. It was not until Palese and colleagues applied the first-generation tools of genetic epidemiology in 1978 that this strain-strain identity could be proven.

Even then, Palese never squarely raised the possibility that the 1977 strain derived directly from a virus stored in a lab freezer for a quarter century. He readily admitted to the discomfort he felt when confronting his RNA mapping results in 1978: "One doesn't want to be accusatory … we don't want to throw stones sitting in the glass house ourselves. … It's difficult to make clear accusations when we really don't know the facts," he said. "I don't think I can get [Chu] into trouble anymore, but I certainly didn't say anything while he was still alive."

Palese told me (DSB) that his 1978 Nature paper was met with scientific curiosity, but no one approached him about the possible biosecurity implications of the identity of the strains. There was no cover-up, just a lack of awareness. Similarly, neither my 2004 Asian Wall Street Journal Op-Ed nor my 2009 New England Journal of Medicine review article on the history of H1N1 viruses stirred serious concerns about biosafety. Even Palese's 2004 declaration in Nature Medicine that the 1977 H1N1 had originated in vaccine trials in the Chinese military, which he attributed to Chu, evoked no biosecurity response. When Rozo and Gronvall (2015) used modern sequencing technologies to reestablish the identity of the 1950 freezer virus and the 1977 pandemic virus, it prompted no reexaminations to explain the identity of the viruses. Only after the emergence of the SARS-CoV-2 pandemic, and the controversy about its possible origin from a lab in Wuhan, China, did these earlier papers on the possible laboratory origins of the 1977 influenza pandemic attract attention.

Lessons Learned

The first and most obvious lesson from the 1977 influenza pandemic is that stored extinct human-adapted viruses are profoundly dangerous. Smallpox is a clear example. The US National Academies have reexamined what justifications there may be for continued R&D using live virulent smallpox (Arvin and Patel 2009; DSB served on this committee). Other than smallpox, the only human pandemic viruses that have gone extinct are the wild polio types 2 and 3, which were driven to extinction by the Global Polio Eradication Initiative, and the early pandemic influenza viruses (H0 and H2, in addition to H1), which were naturally displaced by the emergence of other influenza types. With the passage of time, the proportion of the population susceptible to these now-extinct viruses will continue to grow. SARS-CoV-1 never quite achieved full-blown pandemic status, but it came awfully close, and the world's population remains susceptible. Should any of these proven human-adapted but now extinct viruses be reintroduced into the human population, a pandemic could easily ensue.

The second key lesson from the resurrection and global spread of H1N1 is that "bio-bungling" is a very serious and ever-present threat, comparable to the more widely appreciated threats of biowarfare, bioterrorism, and natural emergence. In bio-bungling, there is no malice or intent to harm. Indeed, the greatest risks of bio-bungling occur when urgent actions are being taken for the best of intentions, as happened in 1977.

Third, a future accidental release of a transmissible virus could occur almost anywhere on the planet. According to the most recent WHO survey, entitled "The Global Proliferation of High-Containment Biological Laboratories," there are currently biosafety level 4 (BSL-4) laboratories in two dozen countries scattered across the globe, and BSL-3 labs in at least 50 other countries (Peters 2018). Release from a lab is an ever-present concern. In addition to lab accidents, flaws in risk assessment may also lead to a release. As this story reveals, the 1977 H1N1 Russian flu likely occurred through a well-intended vaccine R&D initiative. Anywhere that extinct human-adapted viruses are stored, they should be guarded as if they posed a threat comparable to nuclear fissionable material. Tightening of bio-safety protocols in a single country, such as the US, will do little to reduce the threat of future lab-origin pandemics unless laboratories in all countries world-wide implement and enforce comparable standards.

There can be little doubt that we should use the full power of modern science to prevent and respond to pandemic threats. Surveillance, virus isolation and characterization, and R&D on diagnostics, drugs, and vaccines are all crucial elements of an effective pandemic preparedness plan. Research on dangerous viruses that are actively circulating among humans can be carried out with appropriate containment protocols. However, research using viruses that are extinct (not actively circulating) yet known to have pandemic transmission in the past should be carried out rarely, if at all. The fourth and perhaps most important lesson of the 1977 H1N1 pandemic is that in our efforts to control microbial threats by resurrecting extinct pandemic viruses we can inadvertently make things worse—much worse.

Could the 1977 H1N1 influenza pandemic have been averted? Yes. It was a self-fulfilling prophecy pandemic, an entirely avoidable catastrophe.