Spillover: Animal Infections and the Next Human Pandemic

by David Quammen

The virus now known as Hendra wasn’t the first of the scary new bugs. It wasn’t the worst. Compared to some others, it seems relatively minor. Its mortal impact, in numerical terms, was small at the start and has remained small; its geographical scope was narrowly local and later episodes haven’t carried it much more widely. It made its debut near Brisbane, Australia, in 1994.

A zoonosis is an animal infection transmissible to humans.

Reid began wondering whether there might be an exotic virus at work, such as the one responsible for African horse sickness (AHS), a disease carried by biting midges in sub-Saharan Africa. AHS virus affects mules, donkeys, and zebras as well as horses, but it hasn’t been reported in Australia, and it isn’t directly contagious from horse to horse. Furthermore, Queensland’s pestiferous midges don’t generally come biting in September, when the weather is cool.

One was hantavirus—which is actually a group of viruses, long known to virologists following outbreaks in Russia, Scandinavia, and elsewhere but newly conspicuous since a year earlier, 1993, when a new hantavirus emerged dramatically and killed ten people around the Four Corners area of the American Southwest. Australia is justifiably wary of exotic diseases invading its borders, and hantavirus in the country would be even worse news (except for horses) than African horse sickness.

So the creature from Hendra was classified and given a name, based on those provisional identifications: equine morbillivirus (EMV). Roughly, horse measles. About the same time, the AAHL researchers tested a sample of tissue that had been taken from Vic Rail’s kidney during his autopsy. That sample also yielded a virus, identical to the virus from the horses, confirming that this equine morbillivirus didn’t afflict only equines. Later, when the degree of its uniqueness became better appreciated, the label “EMV” was dropped and the virus was renamed after its place of emergence: Hendra.

When a pathogen leaps from some nonhuman animal into a person, and succeeds there in establishing itself as an infectious presence, sometimes causing illness or death, the result is a zoonosis.

Smallpox, to take one counterexample, is not a zoonosis. It’s caused by variola virus, which under natural conditions infects only humans. (Laboratory conditions are another matter; the virus has sometimes been inflicted experimentally on nonhuman primates or other animals, usually for vaccine research.) That helps explain why a global campaign mounted by the World Health Organization (WHO) to eradicate smallpox was, as of 1980, successful. Smallpox could be eradicated because that virus, lacking ability to reside and reproduce anywhere but in a human body (or a carefully watched lab animal), couldn’t hide.

The least conspicuous strategy of all is to lurk within what’s called a reservoir host. A reservoir host (some scientists prefer “natural host”) is a living organism that carries the pathogen, harbors it chronically, while suffering little or no illness.

Nearly all zoonotic diseases result from infection by one of six kinds of pathogen: viruses, bacteria, fungi, protists (a group of small, complex creatures such as amoebae, formerly but misleadingly known as protozoans), prions, and worms.

Viruses are the most problematic. They evolve quickly, they are unaffected by antibiotics, they can be elusive, they can be versatile, they can inflict extremely high rates of fatality, and they are fiendishly simple, at least relative to other living or quasi-living creatures.

Meningitis is a term applicable to any inflammation of the membranes that cover the brain and the spinal cord; it might be caused by a bacterium, a virus, even a reaction to a drug, and it might go away as inexplicably as it appeared.

Provisionally, Field and his colleagues dismissed the bird hypothesis, on two counts. First, they were unaware of any other paramyxovirus that spills over from birds into humans. Second, a mammalian reservoir simply seemed more likely, given that the virus infects humans and horses. Similarity of one kind of host animal to another is a significant indicator of the likelihood that a pathogen can make the leap. Bats are mammals, of course. And bats get around. Furthermore, bats famously harbor at least one fearful virus, rabies, although Australia at that time was considered rabies-free. (Many other bat-virus-human connections would be discovered soon afterward, including some in Australia; but at this time, 1996, the link seemed less obvious.) From the meeting, Field took away a new mandate: Look at bats.

The DPI team also tested old samples from flying foxes, which had been archived for more than a dozen years. Again, they found telltale molecular tracks of Hendra. This showed that the bat population had been exposed to Hendra virus long before it struck Vic Rail’s horses. And then, in September 1996, two years after the Rail outbreak, a pregnant grey-headed flying fox got herself snagged on a wire fence. She miscarried twin fetuses and was euthanized. Not only did she test positive for antibodies; she also made possible the first isolation of Hendra virus from a bat. A sample of her uterine fluids yielded live virus, and that virus proved indistinguishable from Hendra as found in horses and humans. It was enough, even within scientific bounds of caution, to identify flying foxes as the “probable” reservoir hosts of Hendra.

The virus that causes foot-and-mouth disease (FMD) belongs to the picornaviruses, the same group that includes poliovirus and some viruses similar to those that cause the common cold. But infection with FMD virus is a rare misfortune in humans, seldom causing worse than a rash on the hands, the feet, or the mouth lining. More frequently and consequentially, it afflicts cloven-hoofed domestic animals such as cattle, sheep, goats, and pigs.

Impacts of FMD differ from one kind of animal to another. Sheep tend to carry the infection without showing symptoms. Cattle suffer openly and pass the virus to one another by direct contact (say, muzzle to muzzle) or vertically (cow to calf) by suckling. Pigs are special: They excrete far more of the virus than other livestock, and over a longer period of time, broadcasting it prodigiously in their respiratory exhalations. They sneeze it, they chuff it, they oink it, they wheeze it and burp it and cough it into the air. One experimental study found that pig breath carried thirty times as much FMD virus as the breath of an infected cow or sheep, and that once airborne it could spread for miles.

An amplifier host is a creature in which a virus or other pathogen replicates—and from which it spews—with extraordinary abundance.

The amplifier host becomes an intermediate link between a reservoir host and some other unfortunate animal, some other sort of victim—a victim requiring higher doses or closer contact before the infection can take hold.

So three of the four principals in this complex interaction—flying foxes, Hendra virus, and people—have probably coexisted in Australia since the Pleistocene era. Horses arrived in January 1788.

To choose one point is a little artificial, but a good candidate would be the emergence of Machupo virus among Bolivian villagers between 1959 and 1963.

Human-caused ecological pressures and disruptions are bringing animal pathogens ever more into contact with human populations, while human technology and behavior are spreading those pathogens ever more widely and quickly. There are three elements to the situation. One: Mankind’s activities are causing the disintegration (a word chosen carefully) of natural ecosystems at a cataclysmic rate.

Two: Those millions of unknown creatures include viruses, bacteria, fungi, protists, and other organisms, many of which are parasitic.

Three: But now the disruption of natural ecosystems seems more and more to be unloosing such microbes into a wider world.

“Spillover” is the term used by disease ecologists (it has a different use for economists) to denote the moment when a pathogen passes from members of one species, as host, into members of another.

The disease in question, once known as Ebola hemorrhagic fever, is now simply called Ebola virus disease. The pattern stretches from 1976 (the first recorded emergence of Ebola virus) to the present, and from one side of the continent (Côte d’Ivoire) to the other (Sudan and Uganda). The four major lineages of virus that showed themselves during those emergence events are collectively known as ebolaviruses.

In the years since 1996, other outbreaks of Ebola virus disease have struck both people and gorillas within the region surrounding Mayibout 2. One area hit hard lies along the Mambili River, just over the Gabon border in northwestern Congo, another zone of dense forest encompassing several villages, a national park, and a recently created reserve known as the Lossi Gorilla Sanctuary.

Our destination was a site known as the Moba Bai complex, a group of natural clearings near the east bank of the upper Mambili, not far from the Lossi sanctuary. A bai in Francophone Africa is a marshy meadow, often featuring a salt lick, and surrounded by forest like a secret garden. In addition to Moba Bai, the namesake of this complex, there were three or four others nearby. Gorillas (and other wildlife) frequent such bais, which are waterlogged and sunny, because of the sodium-rich sedges and asters that grow beneath the open sky.

Any such spillover in the reverse direction—from humans to a nonhuman species—is known as an anthroponosis. The famous mountain gorillas, for instance, have been threatened by anthroponotic infections such as measles, carried by ecotourists who come to dote upon them.

Combined with destruction of their habitat by logging operations, and the hunting of them for bushmeat to be consumed locally or sold into markets, infectious diseases could push western gorillas from their current levels of relative abundance (maybe a hundred thousand in total) to a situation in which small, isolated populations survive tenuously, like the mountain gorillas, or go locally extinct.

For almost four decades, the identity of Ebola’s reservoir host has been one of the darkest little mysteries in the world of infectious disease. That mystery, along with efforts to solve it, dates back to the first recognized emergence of Ebola virus disease, in 1976. Two outbreaks occurred in Africa that year, independently but almost simultaneously: one in the north of Zaire (now the Democratic Republic of the Congo), one in southwestern Sudan (in an area that today lies within the Republic of South Sudan), the two separated by three hundred miles. Although the Sudan situation began slightly earlier, the Zaire event is the more famous, partly because a small waterway there, the Ebola River, eventually gave its name to the virus.

One factor making the hunt for Ebola’s reservoir especially difficult, especially hard to focus, is the transitory nature of the disease within human populations. It disappears entirely for years at a time.

One animal died and, after it tested positive for Reston virus, forty-nine others housed in the same room were euthanized as a precaution. (Most of those, tested posthumously, were negative.) Ten employees who had helped unload and handle the monkeys were also screened for infection, and they also tested negative, but none of them were euthanized.

The village of Mékouka, on the upper Ivindo River in northeastern Gabon, offers an instance. Mékouka was one of the gold camps in which the outbreak of 1994 got its start. Three years later, a medical anthropologist named Barry Hewlett, an American, visited there to learn from the villagers themselves how they had thought about and responded to the outbreak. Many local people told him, using a term from their Bakola language, that this Ebola thing was ezanga, meaning some sort of vampirism or evil spirit.

One early patient, Hewlett learned, was pulled out of the village clinic because his family disbelieved the Ebola diagnosis and preferred relying on a traditional healer. After that patient died at home, unattended by medical personnel and uncured by the healer, things got testy. The healer pronounced that this man had been poisoned by sorcery and that the perpetrator was his older brother, a successful man working in a nearby village. The older brother was a teacher who had “risen” to become a school inspector and didn’t share the good fortune with his family. So again, as with ezanga among the Bakola people in northeastern Gabon, there were jealous animosities underlying the accusations of sorcery.

Another element of the dangerous brew in and around Mbomo was a mystic secret society, La Rose Croix, more familiar (if barely) to you and me as Rosicrucianism. It’s an international organization that has existed for centuries, mostly devoted to esoteric study, but in this part of the Congo it had a bad reputation, akin to sorcery. Four teachers within one nearby village were members, or were thought to be members—and these teachers had been telling children about Ebola virus before the outbreak occurred. That led some traditional healers to suspect that the teachers had advance knowledge—supernatural knowledge—of the outbreak. Something had to be done, yes? On the day before Barry Hewlett and his wife arrived in Mbomo, the four teachers were murdered with machetes while they worked in their crop fields.

conditions while the primary mission was not science but saving lives, even the experts aren’t sure exactly how the virus typically causes death. “We don’t know the mechanism,” Pierre Rollin told me. He could point to liver failure, to kidney failure, to breathing difficulties, to diarrhea, and in the end it often seemed that multiple causes were converging in an unstoppable cascade. Karl Johnson voiced similar uncertainty, but mentioned that the virus “really goes after the immune system,” shutting down production of interferon, a class of proteins essential to immune response, so that “nothing stops the continued replication of the virus.” This idea of immune suppression by ebolaviruses has also appeared lately in the literature, along with speculation that it might allow catastrophic overgrowth of a patient’s natural populations of bacteria, normally resident in the gut and elsewhere, as well as unhindered replication of the virus itself.

They found evidence suggesting that the medical outcome for an individual patient—to survive and recover, or to die—might be related not to the size of the infectious dose of Ebola virus but to whether the patient’s blood cells produced antibodies promptly in response to infection. If they didn’t, why not? Was it because the virus itself somehow quickly decommissioned their immune systems, interrupting the normal sequence of molecular interactions involved in antibody production? Does the virus kill people (as is now widely supposed) by creating immune dysfunction before overwhelming them with viral replication, which then inflicts further devastating effects?

Ambiguous or not, these results seemed dramatic when they appeared in a paper by Leroy and his colleagues in late 2005. It was a brief communication, barely more than a page, but published by Nature, one of the world’s most august scientific journals. The headline ran: FRUIT BATS AS RESERVOIRS OF EBOLA VIRUS. The text itself, more carefully tentative, said that bats of three species “may be acting as a reservoir” of the virus.

The particle view of Ebola sees it as a relatively old and ubiquitous virus in Central African forests, and each human outbreak as an independent event, primarily explicable by an immediate cause.

The wave view suggests that Ebola has not been present throughout Central Africa for a long time—that, on the contrary, it’s a rather new virus, descended from some viral ancestor, perhaps in the Yambuku area, and come lately to other sites where it has emerged.

What Walsh suggested—to recapitulate in simplest form—is a wave of Ebola virus sweeping across Central Africa by newly infecting some reservoir host or hosts. From its recent establishment in the host, according to Walsh, the virus spilled over, here and there, into ape and human populations. The result of that process is manifest as a sequence of human outbreaks coinciding with clusters of dead chimps and gorillas—almost as though the virus were sweeping through ape populations across Central Africa. Walsh insisted during our Libreville chat, though, that he had never proposed a continental wave of dying gorillas, one group infecting another. His wave of Ebola, he explained, has been traveling mainly through the reservoir populations, not through the apes.

Leroy, on the other hand, has presented his particle hypothesis of “multiple independent introductions” as a diametric alternative not to Walsh’s idea as here stated but to the notion of a continuous wave among the apes.

The four coauthors did agree on a couple other basic points. First, fruit bats might be reservoirs of Ebola virus but not necessarily the only reservoirs. Maybe another animal is involved—a more ancient reservoir, long since adapted to the virus. (If so, where is that creature hiding?) Second, they agreed that too many people have died of Ebola virus disease, but not nearly so many people as gorillas.

Laveran had detected the important truth that malaria is caused by microbes, not by bad air.

But epidemiologists have recognized that, with measles virus, as with other pathogens, there’s a critical minimum size of the host population, below which it can’t persist indefinitely as an endemic, circulating infection. This is known as the critical community size (CCS), an important parameter in disease dynamics.

Another crucial aspect of measles is that the virus is not zoonotic. If it were—if it circulated also in animals living near or among human communities—then the question of critical community size would be moot. There wouldn’t be any necessary minimum size of the human population, because the virus could always remain present, nearby, in that other source.

Their hypothesis: If bonobos carry a form of P. falciparum that is so similar to what humans carry, those parasites may still be passing back and forth between bonobos and us. In other words, falciparum malaria may be zoonotic—in the strict sense of the word, not just the loose sense. Humans in the forests of DRC might be infected on a regular basis with P. falciparum from the blood of bonobos, and vice versa. Their caveat: If that’s so, the great dream of malaria eradication becomes even less attainable.

In plain talk: The human version is one twig within a gorilla branch, suggesting that it came from a single spillover. That’s one mosquito biting one infected gorilla, becoming a carrier, and then biting one human. By delivering the parasite into a new host, that second bite was enough to account for a zoonosis that still kills more than a half million people each year.

(Borneo has been deforested at a high rate within recent decades, to the point that its forest coverage is now less than 50 percent; meanwhile the island’s human population has grown to about 16 million.

“WHO has received reports of more than 150 new suspected cases of severe acute respiratory syndrome (SARS),

Two aspects of what made SARS so threatening were its degree of infectiousness—especially within contexts of medical care—and its lethality, which was much higher than in familiar forms of pneumonia.