Content uploaded by Michael Greger
Author content
All content in this area was uploaded by Michael Greger on Nov 06, 2014
Content may be subject to copyright.
Critical Reviews in Microbiology, 33:243–299, 2007
Copyright
c
Informa Healthcare USA, Inc.
ISSN: 1040-841X print / 1549-7828 online
DOI: 10.1080/10408410701647594
The Human/Animal Interface: Emergence and Resurgence
of Zoonotic Infectious Diseases
Michael Greger
The Humane Society of the United States, Washington, DC, USA
Emerging infectious diseases, most of which are considered
zoonotic in origin, continue to exact a significant toll on soci-
ety. The origins of major human infectious diseases are reviewed
and the factors underlying disease emergence explored. Anthro-
pogenic changes, largely in land use and agriculture, are im-
plicated in the apparent increased frequency of emergence and
re-emergence of zoonoses in recent decades. Special emphasis is
placed on the pathogen with likely the greatest zoonotic potential,
influenzavirus A.
Keywords Agriculture; Avian Influenza; Borrelia burgdorferi;
Bovine Spongiform Encephalopathy; Bushmeat; Campy-
lobacter; Concentrated Animal Feeding Operations;
Deltaretroviruses; Disease Ecology; Disease Evolu-
tion; Domestic Fowl; Emerging Infectious Diseases;
Escherichia coli O157; Extraintestinal Pathogenic
Escherichia coli;Farm Animals; HIV; Influenza A
Virus Subtype H5N1; Listeria monocytogenes; Multiple
Drug Resistance; Nipah Virus; Pandemics; Pets; Public
Health; Salmonella; SARS Virus; Simian Foamy Virus;
Streptococcus suis; viral Hemorrhagic Fevers; Virulence;
Zoonoses
I. INTRODUCTION
Absent the many other impacts of emerging infectious dis-
eases, an analysis solely of the economic ramifications indicates
the costs may be substantial. Based on net changes in GDP
growth forecasts, official (Fan 2003) and academic (Lee and
McKibbin 2003) estimates of the economic costs of SARS in
Asia exceed 10 billion USD. Extrapolating this regional GDP
drop globally, according to the World Bank’s calculus, an in-
fluenza pandemic could represent a loss of 800 billion USD
(Brahmbhatt 2005), though the financial implications of a pan-
demic may surpass SARS in scope and duration (McKibbin and
Sidorenko 2006).
The 1957 and 1968 influenza pandemics cost an estimated
26.8 billion USD (Lederberg et al. 1992), and highly pathogenic
Received 20 August 2007; accepted 22 August 2007.
The author gratefully acknowledges Gowri Koneswaran for invalu-
able editorial assistance.
Address correspondence to Michael Greger, The Humane Society of
the United States, 2100 L St., N.W., Washington, D.C. 20037. E-mail:
mgh1@cornell.edu
avian influenza (HPAI) expenditures are already in the billions
(McLeod et al. 2005). The United Kingdom lost approximately
6 billion USD due to bovine spongiform encephalopathy (Davis
and Lederberg 2001), and the costs associated with antimicrobial
resistance continue to rise (McGowan 2001). HIV/AIDS may
have cost 500 billion USD so far (Annan 2001). HIV now infects
an estimated 38.6 million people, propelling AIDS to the leading
cause of premature death among people 15 to 59 years of age
worldwide (Merson 2006).
The economic, social, and public health burden of these
diseases begs the question: from where do emerging diseases
emerge?
Most human diseases, old and new, are caused by multi-host
pathogens. In a review of 1,407 species of human pathogenic
organisms, 816 (58%) were broadly classified as zoonotic
(Woolhouse and Gowtage-Sequeria 2005), a term coined by Vir-
chow and defined by the World Health Organization in 1959 to
describe “those diseases and infections (the agents of) which
are naturally transmitted between (other) vertebrate animals and
man” (Mantovani 2001).
The exact proportion of emerging human diseases that have
arisen from nonhuman animals is unknown. However, accord-
ing to the Woolhouse and Gowtage-Sequeria review (2005), of
the 177 of 1,407 human pathogens identified as “emerging,” 130
(73%) were zoonotic. The landmark U.S. Institute of Medicine
(IOM) report Emerging Infections concluded that “[t]he signifi-
cance of zoonoses in the emergence of human infections cannot
be overstated” (Lederberg et al. 1992).
A. The First Epidemiological Transition: Domestic Origins
Although there are likely “heirloom” pathogens with which
we co-evolved throughout the evolutionary chain from Homo
to Homo sapiens (Cockburn 1967), most modern human in-
fectious diseases may have been unknown to our hunter and
gatherer ancestors. From the Pleistocene Epoch until fewer than
11,000 years ago, the human population is thought to have ex-
isted as small, nomadic groups of a size and density that pre-
cluded the existence of short-lived, immunizing, human-specific
infections, a proposal strongly put forth by Burnet (1946) and
later supported by increasingly complex mathematical modeling
(Dobson and Carper 1996).
243
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
244
M. GREGER
Medical anthropologists have identified a series of “epidemi-
ological transitions,” major shifts in human disease, since the
beginning of human evolution six million years ago. The advent
of agriculture arguably triggered the first epidemiological transi-
tion, as population size and density grew to levels ripe for the in-
troduction and maintenance of “crowd diseases.” Communities
planting crops settled and grew in relatively fixed locations. This
likely increased their exposure to their own waste as well as to
reservoirs of disease such as unintentional rodent domesticants
that may have set up permanent habitats proximate to human
dwellings (Armelagos et al. 2005).
Although malaria predates agricultural societies, the tran-
sition may have led to rapid speciation and spread of anthro-
pophilic mosquito vectors (Coluzzi 1999), helping to explain
the expansion of virulent Plasmodium falciparum in Africa (Joy
et al. 2003). Similarly, the domestication of cats roughly co-
incides with the acquisition of efficient oral transmission and
subsequent expansion of Toxoplasma (Su et al. 2003).
The transformation of Yersinia pseudotuberculosis,arela-
tively benign fecal-oral route bacillus, into Yersinia pestis, the
flea-transmitted cause of bubonic plague, may have also been
coincident with the rise of agriculture and subsequent food sup-
ply for rodent hosts (Achtman et al. 1999). Similar changes in
rodent ecology have been identified as risk factors for the reap-
pearance of human cases of plague in recent years (Duplantier
et al. 2005).
Although controversy continues to surround the origin of
many infectious diseases (Pearce-Duvet 2006), many of human-
ity’s most serious seem to have been the result of increasingly
close and frequent contact with a new array of potentially
zoonotic pathogens from animals intentionally domesticated
for human use (Ritvo 2004). The same quality that made
certain animals particularly desirable for domestication—the
herd instinct—may have also selected for species that herd
or flock together in concentrations large enough to harbor
epizootic pathogens capable of rapid, population-wide spread
(Diamond 1997). Early humans may have suffered sporadic
cases of animal-borne diseases, such as anthrax from wild
sheep or tularemia (“rabbit skinner’s disease”) from wild rab-
bits (Weber and Rutala 1999), but the domestication of animals
around 10,000 years ago likely provided the greatest conduit for
zoonotic disease transmission (Wolfe et al. 2007).
Human measles likely arose from a rinderpest-like virus of
sheep and goats (Weiss 2001). Only with the prolonged, intimate
contact of domestication was this virus thought able to make the
species jump (Daszak and Cunningham 2002). Smallpox may
have resulted from camel domestication, with camelpox arising
from a cowpox-like, rodent-borne ancestor (Gubser et al. 2004).
Pertussis may have had an ovine or porcine source (Weiss 2001).
In exchange, domestication presented the opportunity to relay
to both cattle and pigs the tapeworms (Taenia) hominids likely
acquired scavenging bovid carcasses (Hoberg et al. 2001).
Human influenza may have only started about 4,500 years
ago with the domestication of waterfowl (Shortridge 2003b),
and leprosy may have originated in water buffalo (McMichael
2001a). Phylogenetic analysis suggests human rhinovirus may
have come from cattle (Rodrigo and Dopazo 1995). Before do-
mestication, the common cold may have been common only to
them.
Mycobacterium tuberculosis was thought to have derived
from Mycobacterium bovis (Sreevatsan et al. 1997), but more
recent evidence suggests a common ancestor (Brosch et al.
2002). This is consistent with its remarkable ability to persist
for decades in host tissues, suggesting a longer period of co-
adaptation. Indeed, unlike pathogens such as rinderpest virus
and camelpox that, today, pose very little or no human risk (Jezek
1983), tuberculosis mycobacteria continue to present a zoonotic
threat to public health, particularly in enzootic areas of the devel-
oping world with high rates of HIV infection (Ayele et al. 2004).
Worldwide, M. bovis TB, which is clinically indistinguishable
in humans from M. tuberculosis TB, is estimated to account for
3% of all manifestations of human tuberculosis (Cosivi et al.
1998). This underscores the importance of reinforcing milk pas-
teurization, animal TB control programs, and airway protection
for meat industry workers (Thoen et al. 2006), and, perhaps,
the deintensification of captive deer farming (Wilson 2002).
Incidentally zoonotic transmission may be bidirectional with
molecular-typed confirmation of human pulmonary TB trans-
mission to cattle (Ocepek et al. 2005), as well as cases of bovine
infection linked to farm workers with genitourinary tuberculosis
urinating on hay (Collins 2000).
All of these animals—birds, bovines, camels, sheep, and
swine—were domesticated exclusively in Eurasia, which may
explain the unilaterality by which indigenous populations in the
Americas were devastated by diseases such as smallpox brought
by the landing Europeans (Diamond 2002). Entire Amerindian
civilizations may have fallen to diseases born on the farm.
Pleistocene Ice Age extinctions eliminated most of the large
mammalian species in the Western hemisphere. With the disap-
pearance of American camels and horses, for example, indige-
nous populations were left with only animals such as llamas
and guinea pigs to domesticate, neither of which seem to carry
much potential for zoonotic disease. This placed the Ameri-
can indigenous peoples in a position both enviable and vulner-
able: Humans in the Americas had the advantage of living rela-
tively free of animal-borne diseases, yet, in the apparent absence
of prior exposure, resistance, or immunity to livestock-derived
zoonoses, faced grave risk when introduced to such immuno-
logically novel pathogens. In fact, livestock-derived zoonotic
diseases contributed to the loss of as much as 90% of the native
American population (Jones 2004). There were buffalo in the
Americas, for example, but no domesticated buffalo, so there
was presumably no opportunity for a disease like measles to
arise until it was brought by Eurasians. Indeed, it is difficult to
pin down a single disease that instead sailed eastbound across
the Atlantic (Weiss 2003).
Additional infections that may have originated with ani-
mal agriculture continue to be discovered. Helicobacter pylori,
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
245
the bacteria for which the 2005 Nobel Prize in Medicine was
awarded, may have first been transmitted to humans via sheep’s
milk and is now thought to spread horizontally in the human
population (Dore et al. 2001). Roughly half of the world’s pop-
ulation may now be infected with H. pylori, making this peptic
ulcer and stomach cancer-associated infection perhaps the most
common chronic infection afflicting humanity (Suerbaum and
Michetti 2002). The closely related Helicobacter pullorum, a re-
cent addition to the list of zoonotic farm animal bacteria (Stanley
et al. 1994), contaminates a large proportion of retail chicken
meat (Atabay and Corry 1998), and may result in a human food-
borne diarrheal illness (De Groote et al. 2000).
Another newly described fecal pathogen, hepatitis E virus
(HEV), is now recognized as a leading cause of fulminant liver
failure in developing countries (Worm et al. 2002), particularly
in pregnant women, with a case fatality of up to 39% found in
a prospective study of women presenting in their third trimester
with acute HEV hepatitis (Beniwal et al. 2003). Scientists be-
gan to suspect the zoonotic nature of this virus when it was
found rampant in North American commercial pork operations
(Yoo et al. 2001). Direct evidence of cross-species transmission
was obtained in 2003 (Tei et al. 2003). Recent isolation of in-
fectious HEV from commercial pig liver sold in U.S. grocery
stores heightens the public health concern over zoonotic HEV
foodborne transmission (Feagins et al. 2007). Unlike dead-end
diseases like trichinosis transmitted by eating improperly cooked
pork, once a disease like hepatitis E crosses the species line, it
may spread human-to-human via blood transfusion (Boxall et al.
2006) or the fecal-oral route, resulting in large outbreaks from
contaminated water (Myint and Gibbons 2007).
B. The Second Epidemiological Transition: Complacency
The second epidemiological transition arguably began a cen-
tury or so ago in countries where improved nutrition, public
health measures, and medical intervention led to both a decline
in infectious disease (Omran 1971) and what the U.S. Insti-
tute of Medicine report called the “era of complacency” (Davis
and Lederberg 2001). In 1948, the U.S. Secretary of State pro-
nounced that the conquest of all infectious diseases was immi-
nent (Najera 1989). Nineteen years later, victory was declared
by U.S. Surgeon General William Stewart: “The war against
diseases has been won” (Morens et al. 2004).
Though premature, Dr. Stewart’s declaration may have been
understandable. By that time in the United States, polio and
rheumatic fever had been conquered, smallpox nearly eradi-
cated, childhood vaccinations developed, and an arsenal of more
than 25,000 different “miracle drug” antibiotic preparations as-
sembled (Selgelid 2005). A Nobel Prize-winning virologist was
even seduced into the heady optimism. In the 1962 text Nat-
ural History of Infectious Disease, Frank Macfarlane Burnet
co-wrote: “To write about infectious disease is almost to write
of something that has passed into history. . . .[T]he most likely
forecast about the future of infectious disease,” he continued,
“is that it will be very dull” (Burnet and White 1962).
C. The Third Epidemiological Transition:
Emerging Infectious Diseases
The situation soon changed. For the first time since the U.S.
federal government began tracking mortality statistics, the de-
clining domestic trend in infectious disease mortality reversed
in 1981, a shift attributed largely to AIDS deaths (Armstrong
et al. 1999). We may currently be living in the third epidemio-
logical transition, which started approximately 30 years ago, the
emergence (or re-emergence) of infectious diseases (Armelagos
et al. 2005) (though it can be argued that if counting interconti-
nental European expansion, we are currently in the midst of the
fourth) (McMichael 2004).
The concept of “emerging infectious diseases” has changed
from a mere curiosity in the field of medicine to an entire dis-
cipline that has been gaining prominence (Brown 2000a). In
recent decades, previously unknown diseases have surfaced at a
pace unheard of in the recorded annals of medicine (Smolinski
et al. 2003)—more than 30 newly identified human pathogens
in 30 years, most of them newly-discovered zoonotic viruses
(Woolhouse 2002).
Although many of today’s doctors learned in their medi-
cal school textbooks that viruses were species-specific (Clough
2004), we now know interspecies navigation to be common.
Creative “kingdom jumping” speculations of a plant reservoir
for Ebola virus aside (Monath 1999), “novel” viruses, as obli-
gate parasites, tend to come from other animal species almost
by definition (Wain-Hobson and Meyerhans 1999).
The trend of increasing zoonotic virus emergence is expected
to continue (WHO/FAO/OIE 2004), and estimates of the poten-
tial number of zoonotic diseases already run into the thousands
(Beran and Steele 1994). The 2003 follow-up to the 1992 In-
stitute of Medicine report on emerging infections suggests that
without appropriate policies and actions, the future could bring
a “catastrophic storm of microbial threats” (Smolinski et al.
2003).
The first surge in species-barrier breaches likely occurred
with the clustering of zoonotic vectors accompanying the do-
mestication of animals—but that was 10,000 years ago. What
new changes are taking place at the human/animal interface that
may be responsible for this resurgence of zoonotic disease in
recent decades?
D. Anthropogenic Impact
Starting in the last quarter of the 20th century, medicine has
been examining emerging infections within an increasingly eco-
logical framework, although the concept is not new. Two cen-
turies ago, Edward Jenner, the founder of modern vaccines, pro-
posed that the “deviation of man from the state in which he was
originally placed by nature seems to have proved to him a pro-
lific source of diseases” (McMichael 2004). This observation
evidently dates back to the 2nd century, when Plutarch argued
that new classes of diseases followed profound changes in the
waywelive (Drexler 2002).
The distinctions raised in the disease emergence debate
framed by Schrag and Wiener (1995) regarding the relative
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
246
M. GREGER
contribution of changes in host versus pathogen and ecology
versus evolution may be largely artificial. As exemplified by the
emergence of antibiotic-resistant bacteria, disease emergence
may be viewed as an evolutionary response to largely anthro-
pogenic environmental changes (Slingenbergh et al. 2004).
The recent rise in zoonotic disease emergence has been par-
alleled by an apparent increase in non-zoonotic infectious dis-
ease emergence among domesticated animals (Mahy and Brown
2000), wild animals (Daszak et al. 2000), domesticated plants
(Rybicki and Pietersen 1999), and wild plants (Anderson and
May 1986). Palumbi (2001) argues that humans are currently
the world’s greatest evolutionary force.
In 2004, a joint consultation on emerging zoonoses was con-
vened by the World Health Organization, the Food and Agri-
culture Organization of the United Nations, and the World Or-
ganization for Animal Health to elucidate the major drivers of
zoonotic disease emergence (WHO/FAO/OIE 2004). A com-
mon theme of primary risk factors for the emergence and spread
of emerging zoonoses was the increasing demand for animal
protein, associated with the expansion and intensification of an-
imal agriculture, long-distance live animal transport, live animal
markets, bushmeat consumption, and habitat destruction.
II. HABITAT DESTRUCTION
In 1933, wildlife ecology pioneer Aldo Leopold reportedly
wrote that “[t]he real determinants of disease mortality are the
environment and the population,” both of which he felt were be-
ing “doctored daily, for better or for worse, by gun and axe, and
by fire and plow” (Friend et al. 2001). Since Leopold penned
those lines, more than half of the Earth’s tropical forests have
been cleared (Pimm et al. 2001). Based on satellite imaging,
global tropical deforestation continues at an annual loss of up to
2–3% (Patz et al. 2004), with the net rate of tropical forest clear-
ing increasing roughly 10% from the 1980s to 1990s (DeFries
et al. 2002).
Grazing animals for human consumption demands an esti-
mated 0.21 hectares per global capita, comparable with the 0.22
hectacres attributed to timber harvesting (Wackernagel et al.
2002). One hundred thousand years ago, likely all members of
the human race lived in eastern Africa (McMichael 2001b). The
world’s population has grown from fewer than 100 million peo-
ple 3,000 years ago to approximately 1 billion at the turn of the
20th century, more than 6 billion by the end of the century, with
2050 projections approaching 10 billion (Tilman et al. 2001).
Two-thirds of this population increase occurred within the last
50 years (De Souza et al. 2003), leading to major ecological
changes and wildlife habitat reduction implicated in zoonotic
disease emergence (Chomel et al. 2007).
A. Nipah Virus
In 1998, a large outbreak of severe, febrile encephalitis oc-
curred among Malaysians exposed to intensively-farmed pigs
that were sickened by fruit bats displaced by slash-and-burn
deforestation. The disease, named after Sungai Nipah New Vil-
lage from which the virus was first isolated, was spread by in-
fected pigs transported around the country. The emergence of
Nipah virus demonstrates the interplay between multiple eco-
logical risk factors cited in the WHO/FAO/OIE report: habitat
destruction, intensive animal agriculture, and long-distance an-
imal transport.
Globally, the rate of tropical deforestation is highest in Asia
(Mayaux et al. 2005). In 1997 and 1998 more than 10 million
acres of virgin forest burned in Borneo and Sumatra, set ablaze
by humans but exacerbated by a severe El Ni˜no Southern Os-
cillation event drought. The haze and habitat loss are thought to
have caused a mass exodus of “flying fox” fruit bats (Pteropus)
searching for food, resulting in an unprecedented encroachment
on cultivated fruit trees (Chua et al. 2002a). Planting orchards
next to pig farms allows for efficient access to swine manure for
fertilizer, but may also result in porcine exposure to flying fox
saliva from partially eaten fruit or urine, from both of which the
novel paramyxovirus was isolated (Chua et al. 2002b).
Though remaining harmless to fruit bats, the Nipah virus ac-
quired the capacity to cause a severe neurological and respiratory
syndrome in young pigs characterized by a loud, nonproductive,
“barking” cough (Mohd et al. 2000). Displaying an unusual host
promiscuity, the virus caused respiratory distress in other ani-
mals proximate to the pig farm—goats, sheep, dogs, cats, and
horses (Uppal 2000). The subsequent trucking of infected pigs to
five states in Malaysia and into Singapore resulted in 229 human
cases, nearly half of which (48%) ended in fatality (CDC 1999).
A cull of more than 1 million pigs in affected areas effectively
ended the outbreak (Lam 2003).
Importantly, previous sporadic deaths of pig farm workers
exposed to encephalitic pigs had been reported elsewhere in
Malaysia the year earlier (Bunning et al. 2000). It may be no
coincidence that the index farm triggering the large outbreak
was one of the largest hog operations in the country with nearly
30,000 animals (Ludwig et al. 2003). Despite being an Islamic
country, annual pig production had risen significantly in the
decade prior to the outbreak, from less than 3 million head to
more than 5 million (FAO 2007b).
Like other contagious respiratory viruses, Nipah virus is
presumably a density-dependent pathogen, requiring a certain
threshold density of susceptible individuals to spread, persist,
and erupt from within a population (Newman et al. 2005). An-
alyzing the management data from the index farm, Olival and
Daszak (2005) concluded that the population size and stocking
density of the pigs, combined with the continual influx of young,
immunologically na¨ıve pigs, may have facilitated the adaptation
and amplification of the virus.
Nipah virus shares similarities with Hendra virus, another
paramyxovirus, identified in Australia after infection resulted
in the deaths of two people handling horses thought infected
by flying foxes (Field et al. 2001). No evidence of prior Hendra
virus infection was found in a serologic study of wildlife
rehabilitators who had had frequent, sustained, and significant
contact with these bats, however (Selvey et al. 1996). This
suggests a requirement for adaptation and/or amplification in an
intermediate host for these viruses, though a case-control study
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
247
of patients in a resurgence of Nipah virus disease in Bangladesh
starting in 2004 linked infection with the consumption of fresh,
raw, date palm sap that was likely directly fruit-bat contaminated
(Luby et al. 2006). Phylogenetic analysis suggests that Nipah
and Hendra are not new viruses (Gould 1996), but remained
uncharacterized until ecological changes precipitated contact
with susceptible hosts.
B. Hemorrhagic Fevers
The 1992 Institute of Medicine report details how new selec-
tion pressures can give rise to new zoonotic viruses (Lederberg
et al. 1992). Stefan Pattyn, the editor of Ebola Virus Haemor-
rhagic Fever (1978), wrote: “The larger the scale of man-made
environmental changes, the greater must be the probability of
emergence of a zoonosis, old or new.” There are few changes
greater than climate change.
Like Nipah virus, the emergence of the Sin Nombre virus in
New Mexico in 1993 was also linked to changes in host reservoir
food supplies brought on by El Ni˜no events (Glass et al. 2002)
that were thought linked to climate change (Trenberth and Hoar
1996). Global warming may also affect the latitudinal and ele-
vational movement of vectored diseases such as malaria, dengue
fever, and West Nile virus encephalitis (Haines and Patz 2004).
One scenario-based modeling study suggested that by 2100, the
few-degree mean global temperature increase could expand the
zone in which malaria is spread from 45% of the world’s popu-
lation to 60% (Martens et al. 1995). The emissions and loss of
forest cover associated with slash-and-burn deforestation con-
tribute to the problem, especially if then used as grazing land, as
the livestock sector has been estimated to contribute more CO
2
equivalents of greenhouse gases than global transport (Steinfeld
et al. 2006).
As leading cattle-producing nations fought during World War
II, Argentina dramatically expanded its beef industry at the ex-
pense of its forests. The ensuing upsurge in field mouse popu-
lations likely led to the surfacing of the Junin virus in the early
1950s (Hoff and Smith 2000) now placing an estimated 5 million
people are now at risk for Argentine hemorrhagic fever infec-
tion (Enria and Feuillade 1998). Neotropical deforestation for
agricultural use has since been associated with the emergence of
other Biosafety Level 4, rodent-borne, hemorrhagic fever are-
naviruses: Machupo virus in Bolivia, Sabia virus in Brazil, and
Guanarito virus, the cause of Venezuelan hemorrhagic fever
(Charrel and de Lamballerie 2003). Inroads into Afrotropical
rainforests exposed other hemorrhagic fever viruses: Lassa, Rift
Valley Fever, and Ebola. All known hemorrhagic fever viruses
are suspected to be zoonotic in nature (Guenno 1997).
Deforestation may also lead to a surge in other zoonotic
pathogen vectors. Many disease-carrying mosquitoes prefer to
breed in the open along partially cleared forest fringes, rather
than deep in the woodland (Sutherst 2004). A recent study of the
primary vector of Falciparum malaria in the Peruvian Amazon
found that mosquito-biting rates in roadside, deforested settle-
ments were up to 278 times higher than those in settlements
without adjacent deforestation (Vittor et al. 2006). When live-
stock are grazed on the cleared land, they may in turn nourish
and amplify arthropod vectors (Patz and Wolfe 2002).
In India, what was initially thought to be a severe outbreak of
yellow fever in 1957 was in fact a novel disease named after the
recently deforested Kyasanur forest and caused by a tick-borne
flavivirus (Taylor 1997). The cleared land was used for the graz-
ing of cattle, a major host for the tick species (Haemaphysalis
spinigera) that carried the virus out from its simian reservoir
to now cause as many as 1,000 human cases each year (Varma
2001).
C. Lyme Disease
Encroachment into natural habitats may bring humans into
new contact with pathogens, or, in the case of Nipah virus and the
reemergence of vampire bat-borne rabies in the Amazon Basin
(Schneider et al. 2005), encroachment may flush pathogens out
to us. In Connecticut, where Lyme disease was first recog-
nized in 1975, encroachment might be more properly termed
“suburbanization.”
The density of infected, nymphal, black-legged ticks (Ixodes
scapularis), the primary risk factor for Lyme disease, has been
linked to increasing forest fragmentation in periurban areas,
likely due to corresponding increased numbers of the white-
footed mouse (Peromyscus leucopus), the principal reservoir of
the Lyme bacterium (Borrelia burgdorferi) (Allan et al. 2003).
This “edge” habitat effect on P. leucopus density is thought due
to decreased predation and competition in small forest fragments
interspersing residential development (Nupp and Swihart 1996).
Brownstein et al. (2005), though, compared human disease in-
cidence with habitat fragmentation using satellite imagery of
landscape forest patches and showed that entymological risk
does not necessarily translate into increased human risk, per-
haps because, at a certain level of suburban development, fewer
residential properties adjoin woodlots (Maupin et al. 1991).
III. BUSHMEAT
Humans and our ancestors have likely consumed bushmeat,
wild animals killed for food, for millions of years (Boesch
1994). During the 20th century, however, commercial hunting
using firearms and wire snares to supply logging (Wilkie et al.
1992) and oil exploration (Thibault and Blaney 2003) operation
concessions along new roadway networks has dramatically in-
creased the catch in Central African forests (Fa and Garc´ıa-Yuste
2001). Annually, it is estimated that 579 million wild animals are
caught and consumed in the Congo basin, equaling 4.5 million
tons of bushmeat, with the addition of a possible 5 million tons of
wild mammalian meat from the Amazon basin (Fa et al. 2002).
Tropical lowland forest habitat contains the world’s greatest
terrestrial biodiversity (Turner 1996) and may therefore harbor a
reservoir of zoonotic pathogens. Logging in Central Africa gen-
erally involves selective extraction of high-value timber species
rather than clear-cutting, which may maintain this higher natu-
ral density of potential hosts (Fa et al. 1995). Karesh and Cook
(2005) estimate that the wildlife trade in general generates in ex-
cess of one billion direct and indirect contacts between humans
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
248
M. GREGER
and domesticated animals annually. The broad range of tissue
and fluid exposures associated with the bushmeat industry’s
hunting and butchering may make these wildlife interactions
especially risky (Wolfe et al. 2000).
Hunting has been associated with the contraction of diseases
such as contagious pustular dermatitis, a disease caused by the
parapoxvirus Orf virus (Wolfe et al. 2000), and monkeypox,
which is associated with the preparation and consumption of in-
fected rodents (Pattyn 2000). In Africa, as many as 30 different
species of primates are also hunted and processed by the bush-
meat industry (Peeters et al. 2002). Conservation concerns aside
(Walsh et al. 2003), primate butchering and consumption may
present particular risk of cross-species disease transmission. The
hunting of red colobus monkeys, for example, was implicated in
a localized epidemic of human monkeypox that persisted for four
generations of probable human-to-human transmission (Jezek
et al. 1986).
Evolutionary distance may explain why humans are more
likely to get “mad cow disease” than Dutch elm disease. Near-
universal taboos against cannibalism may have well-served the
human species by reducing enteral and parenteral exposure to
well-adapted viruses (Karpowicz et al. 2004), but full sequenc-
ing of the chimpanzee genome has revealed how closely hu-
mans are related to other anthropoid primates (Gunter and Dhand
2005).
By consuming other members of the family Hominidae, hu-
mans may be exposed to pathogens sharing similar host adap-
tations (Brack 1987). For instance, human outbreaks of Ebola
hemorrhagic fever have been traced to exposure to the dead bod-
ies of infected chimpanzees (WHO 1996) and gorillas (WHO
2003) hunted for food. Indeed, the primary risk factor found
for Ebola virus infection among wild chimpanzees was hunting
behavior (Formenty et al. 1999), which involves not only the
consumption of raw meat, but a rudimentary butchering with
mouth and hands (Wolfe et al. 2000). Ebola virus is the cause
of one of humanity’s deadliest infections, but is not efficiently
spread (Legrand et al. 2007) compared to a virus like HIV.
A. HIV
The year the global eradication of smallpox was declared
(WHO 1980b), the first reported cases of AIDS were being iden-
tified (CDC 1981). Since then, HIV/AIDS has resulted in an es-
timated 65 million infections and 25 million deaths. In several
African countries, the estimated prevalence of HIV now exceeds
20% of the adult population (UNAIDS 2006).
Although blood banking, the relaxation of sexual mores, and
injection drug use facilitated the spread of HIV, the simplest
and most plausible explanation for the emergence of the virus
appears to be exposure to animal blood or excretions as a result of
hunting and butchering primates, or the subsequent consumption
of uncooked or contaminated bushmeat (Sharp et al. 2001).
Blood samples obtained from 573 freshly butchered primates
in logging concessions and bushmeat markets found 18.4% with
evidence of Simian Immunodeficiency Virus (SIV), considered
the precursor to HIV (Peeters et al. 2002). Although HIV-like
viruses have been recently discovered in gorillas, the butchering
of chimpanzees is considered the most likely source for HIV-1,
the strain of the AIDS virus that has spread around the world
(Van Heuverswyn and Peeters 2007).
Molecular analyses now suggest SIV crossed species not
once, but on no fewer than eight separate occasions in recent
history (Peeters et al. 2002). Until a few decades ago, such a
zoonotic transmission event may have only affected a small num-
ber of isolated rural villages. Corporate logging for export ex-
panded western Africa’s tradition of primate subsistence hunting
(Hahn et al. 2000) into a major commercial enterprise (Larkin
2000) that extends into surrounding major cities where ape meat
garners premium prices in restaurants (Weiss and Wrangham
1999). Combined with reports of the frequent rotation of sex
workers into logging camps by timber companies (Nisbett and
Monath 2001), the routes of transmission along roads cut by
deforestation in Africa may run both ways (Peeters 2004).
In addition to concurrent socioeconomic disruptions of post-
colonial, sub-Saharan African infrastructure, widespread iatro-
genic and self-injection use of unsterile needles may have played
a role in the cross-species adaptation of the virus and account
in part for the simultaneous appearance of multiple strains of
HIV across Africa (Marx et al. 2001). Similar possible iatro-
genic facilitation of infection has been reported with parenteral,
antischistosomal vaccination in Egypt, which may have led to
widespread hepatitis C virus distribution (Frank et al. 2000), as
well as the exposure of millions of Americans to simian virus
40 in batches of polio vaccine between 1955 and 1963 (Stratton
et al. 2002).
While zoonotic diseases such as rabies may result in the
deaths of approximately 50,000 people globally every year
(Meslin et al. 2000), for most zoonoses, humans are the dead-
end host (Woolhouse and Gowtage-Sequeria 2005). In terms
of global public health implications, the greatest concern sur-
rounds zoonotic infections like HIV/AIDS that can not only
jump from animals to humans, but can then spread human-to-
human. More people engage in sexual activity than the butcher-
ing of chimpanzees.
Emerging retroviruses are of special concern because of
their ability to insinuate into host DNA. There are seven gen-
era in the Retroviridae family: lentivirus, including HIV and
SIV, spumavirus, and five groups of cancer-causing retroviruses
previously clustered together as oncovirus (Knipe and Howley
2007). SIV is just one of a large reservoir of poorly characterized
lentiviruses in African primates raising the specter of additional
AIDS-like zoonoses arising from continued bushmeat consump-
tion (Peeters et al. 2002).
Retroviruses are also highly recombinogenic (Hu and Temin
1990). Given the high prevalence of HIV in Central Africa
(UNAIDS 2006), a vulnerable pool of immunosuppressed hosts
might not only facilitate the establishment of novel pathogens,
but act as a retroviral mixing vessel for the recombination of
HIV with other primate lentiviruses. This could potentially re-
sult in changes in HIV drug resistance, tropism, or virulence
(Chen et al. 2006b).
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
249
B. Other Primate Retroviruses
1. SFV
In 1971, a spumavirus was isolated from a nasopharyngeal
carcinoma in a Kenyan patient (Achong et al. 1971) sharing 95–
97% amino acid homology with a strain found in chimpanzees
whose natural habitat includes Kenya (Schweizer and Neumann-
Haefelin 1995). This suggested zoonotic transmission, but labo-
ratory contamination could not be ruled out (Meiering and Linial
2001). Subsequent case reports of laboratory workers exposed to
primate fluids have shown evidence of persistent infection with
the spumavirus simian foamy virus (SFV), confirming human
susceptibility to other classes of primate retroviruses (Heneine
et al. 1998).
Cameroon is representative of a region in Africa in which
a noteworthy array of new zoonotic pathogens has been iden-
tified, including HIV/AIDS, Ebola and Marburg viruses, and
monkeypox. The country is home to among the highest density
of large anthropoid primates in Africa, as well as high human-
exposure frequency to wild primate tissues from hunting (Wolfe
et al. 2000). Indeed, in one survey of rural villagers involved in a
community-based HIV-prevention campaign in Cameroon, 61%
of participants reported direct exposure to fresh, non-human pri-
mate blood and body fluids (Wolfe et al. 2004), making this
area well-suited for a study of the zoonotic potential of SFV in
bushmeat.
Ten out of 1,099 (1%) Cameroonese adults sampled had anti-
bodies to SFV, suggesting that primate retroviral zoonoses were
more frequent, widespread, and contemporary than previously
appreciated (Wolfe et al. 2004). Road networks now connect-
ing such remote populations with urban access to international
travel could allow for global dissemination of any emerging
bushmeat-borne zoonoses (Robbins et al. 2003).
Although SFV infection is enzootic in many wild pri-
mate populations, seroprevalence is generally higher in pri-
mates housed in captivity, approaching 100% in some studies
(Meiering and Linial 2001). Schweizer et al. (1995), for exam-
ple, found 93% of captive African green monkeys infected with
SFV, compared to a separate study in which 36% of wild mon-
keys tested positive (Stiles 1968). Given this propensity, animal
handlers from primate research centers and zoos in the United
States were serologically screened to further gauge the extent of
SFV spread.
Tennew SFV-infected persons were identified out of the 187
lab and zoo workers studied (5.3%) (Switzer et al. 2004). Since
few of the primate contacts involved blood exchanges, saliva was
considered the likely route of transmission. A later study found
that entertainment workers could also be at risk of infection, as
SFV was detected by PCR in 52.9% of a small sample of per-
forming macaques in Indonesia (Schillaci et al. 2005). Primate
pet owners and tourists visiting “monkey temples” in Asia also
risk exposure (Jones-Engel et al. 2006).
The persistence of infection is characteristic of foamy viruses,
with recovery of infectious SFV from a human host demon-
strated 20 years after zoonotic transmission, despite high anti-
body titers (Schweizer et al. 1997). Putative foamy virus dis-
ease associations in humans have been made with dialysis en-
cephalopathy (Cameron et al. 1978), de Quervain’ thyroiditis
(Werner and Gelderblom 1979), multiple sclerosis (Westarp
et al. 1993), sensorineural hearing loss (Pyykko et al. 1994), fa-
milial Mediterranean fever (Tamura and Kira 1995), myasthenia
gravis (Liu et al. 1996), and Graves’ disease (Lee et al. 1998),
butnodefinitive links have been found (Meiering and Linial
2001).
Two AIDS patients dying after baboon liver transplants
showed PCR evidence of baboon foamy virus sequences in tis-
sues distant to the transplant (Allan et al. 1998), but infection of
human cells was not specifically analyzed, raising the possibility
that the viral sequences found were due to migrating donor cells
(i.e., leukocytes). Without a study of SFV infection in immuno-
compromised hosts, supposition that SFV has a more aggressive
course in an immunologically permissive environment will re-
main speculative (Jones-Engel et al. 2005).
Spouses of SFV-exposed animal handlers show no evidence
of infection, suggesting the virus is not readily spread human-to-
human (Switzer et al. 2004). More recently, however, SFV in-
fection was described in a Central African sex worker (Switzer
et al. 2005). The prevalence of SFV infection among lab and
zoo workers has raised concerns about blood supply safety. The
demonstration of SFV transmission via experimental whole-
blood transfusions in primates has led to the consideration of
excluding those with ongoing occupational primate exposure
from the blood donor pool (Brooks et al. 2007).
2. HTLV
Further analysis of the Cameroon blood samples also re-
vealed evidence of emerging zoonotic transmission of a third
class of human retroviruses, deltaretrovirus (Wolfe et al. 2005).
Two strains of human T-lymphotropic virus (HTLV), types 1 and
2, were already known to infect an estimated 15 to 20 million
people globally, spread sexually, vertically, or parenterally, via
blood transfusion or IV drug use (Mahieux and Gessain 2003).
Also known as human T-cell leukemia/lymphoma virus, HTLV-
1 can have a long latency period before triggering an aggressive
hematological malignancy in 1 in 1,500 adult HTLV-1 carriers
per year (CDC 1987). Both HTLV-1 and HTLV-2 are also asso-
ciated with a progressive neurodegenerative disease known as
tropical spastic paraparesis (Roucoux and Murphy 2004) and are
thought to have an ancient zoonotic primate source (Courgnaud
et al. 2004).
Wolfe et al. (2005) first identified HTLV-3 and HTLV-4 in
Cameroon monkey and ape hunters. These findings suggest that
retroviruses continue to cross actively into human populations,
though whether HTLV-3 and HTLV-4 are pathogenic or can be
passed between humans remains to be shown (Kuehn 2005).
Because the latency period of viruses like HTLV can be
decades (Blatter 1989), there is concern that emerging primate
retroviruses could circulate in the human population years be-
fore discovery, at which point it may be too late to stop an epi-
demic (Fricker 2000). Although improved global public health
surveillance and blood supply screening may aid early detection,
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
250
M. GREGER
efforts to curtail the commercial bushmeat trade might reduce
the frequency of nascent zoonotic transmission.
C. SARS
The expansion of the bushmeat trade in Africa has been
well-documented (Barnes 2002), but a similar surge in Asia
has received less attention (Bennett and Rao 2002). The inten-
sive commercial bushmeat trade likely got its start in Asia (Bell
et al. 2004), particularly in Guangdong, the southern Chinese
province surrounding Hong Kong SAR and presumed birthplace
of both the highly pathogenic avian influenzavirus A subtype
H5N1 (H5N1) (Chen et al. 2004) and the epidemic of atypical
pneumonia referred to as severe acute respiratory syndrome, or
SARS (Donnelly et al. 2003).
Anovel coronavirus, SARS-associated coronavirus (SARS-
CoV), was identified (Peiris et al. 2003). Before SARS, Coro-
naviridae were typically only known to cause a mild, com-
mon cold-like illness in humans (Siddell 1995), though they
are a common occurrence in a range of avian and mammalian
species and include turkey, feline, and canine enteric coron-
aviruses; infectious bronchitis virus in chickens; porcine hemag-
glutinating encephalomyelitis virus; porcine and bovine enteric
and respiratory viruses; and murine hepatitis virus (Holmes
1999).
The preponderance of persons handling, killing, or selling
caged, wild animals used for food among early index SARS
cases in Guangdong (Xu et al. 2004), combined with the de-
tection of SARS-CoV-like viruses in both animals and workers
at Chinese live animal markets, suggested the market system
played a role in the amplification and transmission of the virus
(Guan et al. 2003). A subsequent survey of animals in retail
markets focused attention on the cat-like masked palm civet
(Paguma larvata) (Guan et al. 2003).
In China, animals are eaten for enjoyment, sustenance, and
for their purported medicinal qualities (Lawrie 2004). In addi-
tion to civets being trapped and raised for meat, civet penis is
consumed as an aphrodisiac (Bell et al. 2004) and coffee beans
fed to captive civets are recovered from their excreta, flavored
with perianal gland scent (William 2003).
The search for the original zoonotic reservoir of SARS has
narrowly honed in on wildlife farms, animal markets, and restau-
rants in China without recognition that many of the wild animals
sold are smuggled into the country via an expanding regional
black market trade (Bell et al. 2004). Roberton et al. (2003) doc-
umented a vast expansion of the wildlife trade in the 1990s, in
which the commercial sale of exotic species to supply the emerg-
ing urban middle-class demand largely supplanted subsistence
hunting. Many of the wild animals, typically while still alive,
enter China through Vietnam from Lao PDR where the bush-
meat trade rose to become the second-largest income source for
rural families (Duckworth et al. 1999). Studying a single Lao-
tian market, Srikosamatara et al. (1992) estimated the per-annum
sale of 8,000 to 10,000 animals of at least 23 mammalian species,
6,000 to 7,000 animals of more than 33 avian species, and 3,000
to 4,000 reptiles, for a total weight of 33,000 kilograms.
With growing populations and increasing demands for
wildlife meat outstripping local supplies, the transboundary
wildlife trade has been supplemented with the creation of in-
tensive, captive production farms in which wild animals may
be raised under poor sanitation in unnatural stocking densities
(Karesh et al. 2005) before being transported and caged at mar-
kets for sale. Selling animals while still alive guarantees fresh-
ness in the minds of many regional consumers, but may result
in potential pathogen discharge into the retail setting (Brown
2004). Together, these factors have been implicated in the am-
plification of emerging zoonotic pathogens with the potential to
infect hunters, farmers, and market patrons (Gilbert et al. 2004).
Seroprevalence studies have found that while civets at farms
that supply the live markets were largely free of infection, up
to 80% of those at markets harbored significant antibody titers
to SARS-CoV (Tu et al. 2004). This suggests that most infec-
tions happened at the market, perhaps because of a combination
of crowded interspecies mixing and the immunosuppressive ef-
fect of stress (Padgett and Glaser 2003). Following the SARS
outbreak, the Chinese government reportedly confiscated more
than 800,000 wild animals from the markets of Guangdong.
Recent data suggest that the masked palm civet is merely an
intermediate host transmitting SARS-CoV-like viruses to hu-
mans from their natural host reservoir, bats (Li et al. 2005),
which are also eaten in Guangdong Province (Woo et al. 2006).
Just as bats are likely the original source of all terrestrial car-
nivore rabies (Lyssavirus) viruses (Bardrane and Tordo 2001),
they may also be the starting point for all coronavirus lineages
(Vijaykrishna et al. 2007).
Analogous to the role pigs played in Nipah virus disease,
civets may facilitate transmission of SARS-CoV-like viruses by
acting as an amplifying host with frequent human contact, but
may be more than just passive conduits. Genomic sequencing
of SARS-CoV-like viruses in palm civets over time suggests not
only viral adaptation (increased viral loads) to the civets (Song
et al. 2005), but perhaps adaptation to humans as well.
The SARS-CoV spike (S) protein has been shown to be re-
sponsible for virus attachment to susceptible cells (Hofmann
et al. 2004), and two key amino acid residues were identified
as critical for species-specific receptor binding (Qu et al. 2005).
Using pseudotyped lentiviruses, Qu et al. (2005) demonstrated
that substituting both residues in the human SARS virus with
those found in civet isolates nearly abolished human cell infec-
tivity, and single substitutions reduced receptor binding affini-
ties. Conversely, the reverse double substitution in the civet virus
enabled it to infect human cells.
After the initial SARS epidemic ended in July 2003, four new
human cases were confirmed in Guangdong that winter. Un-
like most of the previous cases, they had not had close contact
with infected persons and presented with mild clinical symp-
toms. Epidemiologic and phylogenetic analyses indicated that
these cases likely represented a second, independent zoonotic
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
251
transfer. Viral isolates from palm civets at both a local market
and the restaurant with which three of the four human cases were
associated were sampled at the time and were nearly identical
to those found in the new, milder human cases. Civet isolates
were detected that shared one of the two civet-to-human S pro-
tein critical residue substitutions found in all of the new human
patients and none of the prior year’s civet isolates. The findings
suggested a cross-host “exaptive” adaptation to civets may have
resulted in greater human infectivity (Song et al. 2005).
Following the reemergence of SARS cases, Kan et al. (2005)
conducted a more comprehensive phylogenetic analysis of
SARS-CoV-like viruses in market civets and concluded that
the evolutionary starting point for SARS-CoV was a prototype
group of viruses exclusive to nonhuman animals that gave rise to
the low pathogenicity group responsible for the later sporadic,
mild human cases, which in turn led to a highly pathogenic
grouping of viruses responsible for the 2003 epidemic. Based
on this molecular epidemiology data, Zhao (2007) suggested
that the palm civet may serve not only as transitional host and
amplifier of SARS-CoV-like viruses, but a “favorable incubator”
as well.
Interestingly, genomic and molecular clock analyses suggest
that the bovine coronavirus now causing shipping fever disease
jumped species around the year 1890, resulting in the emergence
of human coronavirus OC43 (Vijgen et al. 2005). This tempted
Vijgen et al. to conjecture that the 1889-1890 pandemic ascribed
to influenza may have been the result of a SARS-like interspecies
transmission event triggered by the reported mass culling of
cattle herds between 1870 and 1890, an extensive depopulation
associated with a highly pathogenic bovine respiratory disease
panzootic.
IV. LIVE ANIMAL TRANSPORT
Culinary choice was likely not the only key factor in the emer-
gence and spread of SARS-CoV. Like HIV, historical zoonotic
transmission and adaptation of SARS-CoV-like coronoviruses
may have died out in a remote rural setting. Instead, given today’s
unprecedented speed, volume, and reach of global air travel, a
worldwide outbreak of SARS was seeded from a single person
on a single day on a single floor of a Hong Kong hotel (MacKen-
zie et al. 2004). That index case, a physician from Guangdong,
had attended a wedding. As guests departed, the virus coughed
by one man spread to five countries within 24 hours (Lee and
Krilov 2005). Within months, the virus spread to 30 countries on
six continents (Mack 2005), causing 8,096 probable cases and
774 deaths (WHO 2004). In the past, a trip around the world took
a year; today, we and our viral baggage can circle the globe in
24 hours (Johnson 2003). The annual number of globe-trotters
now exceeds an estimated 1 billion people (Lashley 2004).
A. Exotic Pet Trade
Humans are not the only global travelers, and zoonotic vec-
tors do not make their way out of jungles solely as bushmeat.
Indeed, the first filovirus discovery was not in Africa, but 4,000
miles away in Marburg, Germany, causing an outbreak that had
its origins in a laboratory that received monkeys from Uganda
(Martini 1969). Likewise, Ebola virus, from the same family
Filoviridae,was imported into the United States in monkeys des-
tined for research (Jahrling et al. 1990). Holding different pop-
ulations of captive primates together during transport or while
housed at or en route to zoos or laboratories might result in
viral recombination, potentially resulting in unpredictable, new
strains that could otherwise never have occurred (Chapman et al.
1995).
Animals translocated for other human uses have also been
implicated in the spread of zoonotic disease. The international
movement of horses for equestrian competitions has been asso-
ciated with the spread of equine influenza (Powell et al. 1995;
Guthrie et al. 1999). After their introduction into Eastern Eu-
rope and New Zealand, fur-bearing animals became new reser-
voirs for rabies (Gylyset al. 1998) and M. Bovis (Woodford and
Rossiter 1993), respectively. Rabies was ferried into the mid-
Atlantic United States in the 1970s when hunting pens were re-
populated with raccoons trapped in enzootic zones of the South
(Woodford and Rossiter 1993), and the movement of hares from
central and Eastern Europe for sporting purposes has led to sev-
eral outbreaks of tularemia (Godfroid et al. 2005). Swapping the
wings of Aedes aegypti for those of a jumbo jet, the dengue fla-
vivirus has triggered major intercontinental outbreaks of dengue
fever traced to inadvertent live animal transport of mosquitoes
and their eggs (Reiter and Sprenger 1987).
Whether for exotic cuisine or exotic pets, animals imported
together may end up in holding areas in dealer warehouses,
where they and their viruses can further interact. The first human
outbreak of monkeypox in the Western hemisphere, resulting in
71 cases across six Midwestern U.S. states (Gross 2003), was
traced to monkeypox-infected Gambian giant rats shipped to a
Texas animal distributor along with 800 other small mammals
snared in Ghana. Notably, there were no infections recorded di-
rectly from the African rodents; exposure was derived from sec-
ondary and tertiary contacts often inherent to the trade (Chomel
et al. 2007).
A similar situation exemplifying the multiplicity of inter-
species contacts was thought to have transpired in 2005 when
the British government announced that H5N1 virus had been
detected in an imported parrot. This followed a case the year be-
fore of a Thai man stopped for a routine random drug check in
a Belgian airport and discovered to be smugglinga pair of rare
crested hawk eagles stuffed into plastic tubes in his luggage.
Both of the birds were found to be H5N1-infected (Borm et al.
2005). The parrot, however, had been imported from H5N1-free
Suriname. The working hypothesis was that the bird contracted
the virus while housed in a quarantine facility with finches from
Taiwan found to be infected, though Taiwan had similarly been
declared bird flu-free. Perhaps infected birds had been smug-
gled from China into Taiwan for global export, infected the Tai-
wanese birds, which were then legally imported into the U.K. to
mix with and infect the South American parrot (DEFRA 2005).
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
252
M. GREGER
(The British Government later admitted a procedural laboratory
error in which the parrot and finch samples were accidentally
pooled, calling the psittacine infection into question (Redrobe
2007), but the mistake still does not explain how the finches
had first become infected.) Indeed, that same month, a freighter
reportedly had been caught trying to smuggle into Taiwan from
mainland China more than 1,000 birds, some infected with H5N1
(COA 2003).
Pet bird shipments have also been implicated in outbreaks of
psittacosis (Moroney et al. 1998) and their involvement specu-
lated in the introduction of West Nile virus from the Middle East
(Lanciotti et al. 1999) into the Western hemisphere (Rappole
et al. 2000). Herpes B virus is enzootic in macaques (Weigler
1992) yet is a rare though often fatal cause of human fulminat-
ing meningoencephalitis (Holmes et al. 1995), which, combined
with reports of rabies exposure (Favoretto et al. 2001), raises
concerns about people keeping primates as pets (Ostrowski et al.
1998). African pygmy hedgehogs have triggered salmonellosis
in the Americas (Riley and Chomel 2005), and terrapins shipped
from the Americas have caused salmonellosis in Ireland (Lynch
et al. 1999).
Worldwide, an estimated 40,000 primates, 4 million birds,
640,000 reptiles, and 350 million tropical fish are traded live each
year for the pet trade (Karesh et al. 2005). Wildlife trafficking
in exotics is estimated to be a 6 billion USD industry (Check
2004).
B. Livestock Transport
Over the last century, there has been a shift away from live-
stock production as a highly localized enterprise, where animals
were typically born, fattened, and slaughtered in the same re-
gion. The number of live animals traded for food quintupled
in the 1990s, with more than one billion moved across borders
in 2005 (FAO 2007b). Long-distance live animal transport may
make countries more vulnerable to acts of bioterrorism with
zoonotic agents (GAO 2005), a risk thought amplified by the
concentrated and intensive nature of contemporary industrial
farming practices (Chalk 2004). The transport and crowding of
animals from different herds or flocks in poorly ventilated and
stressful environments for long periods has been considered ide-
ally suited for spreading disease (FAO 2002).
The immunosuppressive stress of prolonged live animal
transport may not only increase a healthy animal’s susceptibility
to infection, but may trigger the emergence of a variety of diar-
rheal and respiratory diseases caused by endogenous microogan-
isms that might not normally lead to disease. So-called “ship-
ping fever,” for example, the bovine version of which costs U.S.
producers more than 500 million USD a year, is often caused
by latent pathogens that may become active when transporting
cattle long distances (Roth and Flaming 1990). Approximately
3,000 truckloads of cattle are moved per day in the United States
(Crews 2004).
Long-distance transport may also increase the fecal shedding
of zoonotic disease agents. Barham et al. (2002) found the av-
erage prevalence of Salmonella within feces and on the hides of
cattle to be 18% and 6%, respectively, before transport. After
the animals were loaded onto a vehicle and trucked for 30 to 40
minutes, Salmonella levels found in feces increased from 18%
to 46%, and the number of animals with contaminated hides es-
calated from 6% to 89% upon arrival at the slaughter plant. A
recent study found that from the time cattle were loaded onto
tractor-trailers at the feedlot to the time their hides were removed
during processing, the prevalence of the enterohemorrhagic E.
coli O157:H7 on hides increased from 50.3% to 94.4% (Arthur
et al. 2007). E. coli O157:H7 isolates from transport trailer walls
have been genotypically matched to hide samples in the packing
plant (Childs et al. 2006). Fecal pathogens on the hide may then
end up in the meat supply (Arthur et al. 2007). Similar results
were found in pigs (Marg et al. 2001) and chickens (Whyte et al.
2001) raised and transported for meat.
Regarding the spread of the Nipah virus in Malaysia in 1999,
the Thai Minister of Public Health was quoted as saying: “A
hundred years ago, the Nipah virus would have simply emerged
and died out; instead it was transmitted to pigs and amplified.
With modern agriculture, the pigs are transported long distances
to slaughter. And the virus goes with them” (Specter 2005). One
hundred people died in the Malaysian outbreak (CDC 1999),
whereas avian influenza viruses are thought to have the potential
to spawn pandemics capable of killing millions (Murray et al.
2006).
In early 2004, outbreaks of the highly pathogenic avian in-
fluenza virus H5N1 were reported across eight countries in
Southeast Asia nearly simultaneously. Given the pattern and tim-
ing of outbreaks, the Food and Agriculture Organization (FAO)
of the United Nations identified the transport of live birds reared
for human consumption as a primary culprit in the rapid spread
(FAO 2007a). The FAO reported, for example, that 5,000 chick-
ens succumbed to avian influenza in Lhasa, Tibet, and that these
infected birds were transported to Tibet from China’s Lanzhou
city—a trip spanning more than 1,600 kilometers (FAO 2004b).
The further the animals are transported, the further the diseases
may spread (DG-SANCO 2002).
Long-distance live animal transport has also been blamed
for the spread of swine influenza viruses in the United States,
where livestock may travel an average of 1,000 miles (Wilson
et al. 2000). Throughout much of the 20th century, influenza
viruses had established a stable H1N1 lineage within U.S. pigs,
becoming one of the most common causes of respiratory disease
on North American pig farms (Zhou et al. 1999). That seemed to
have changed in August 1998 when thousands of breeding sows
fell ill on a North Carolina pig farm. An aggressive H3N2 virus
was recovered, bearing the H3 and N2 antigens of the human
influenza strain circulating since 1968. Not only was this highly
unusual, but, upon sequencing of the viral genome, researchers
found that it was not just a double reassortment (a hybrid of hu-
man and pig viruses, for example), but a never-before-described
triple reassortment of human, avian, and porcine influenza virus
gene segments (Zhou et al. 1999), which raised concerns about
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
253
further mammalian adaptation of influenza virus (Wuethrich
2003b).
The continuous cycle of U.S. mass animal movement, involv-
ing inter-auction movements and intra-auction mixing, may pro-
vide a built-in dispersal mechanism for potentially zoonotic dis-
ease agents (Shields and Matthews 2003). Within months of the
appearance of the new swine flu virus in North Carolina, it sur-
faced in Texas, Minnesota, and Iowa (Zhou et al. 1999). Within
one year, it had spread across the United States (Webby et al.
2000). The rapid dissemination across the nation was blamed on
the cross-country transport of live pigs (Wuethrich 2003b). In the
United States, pigs travel coast to coast, frequently born in North
Carolina, fattened in the corn belt of Iowa, and then slaughtered
in California. It is often cheaper to transport the animals to the
feed rather than transport the feed to the animals (USDA ERS
2003). While this regional segmentation of production stages
may cut down on short-term costs for the pork industry (Bur-
rell 2002), the highly contagious nature of zoonotic diseases
like influenza, perhaps made further infectious by the stresses
of transport (Wuethrich 2003b), must be considered when cal-
culating the true cost of long-distance live animal transport.
V. LIVESTOCK REVOLUTION
Driven by the population explosion, urbanization, and in-
creasing incomes, the per-capita consumption of meat, eggs,
and dairy products has dramatically expanded in the developing
world, leading to what has been termed the “Livestock Revolu-
tion” beginning in the 1970s, akin to the 1960s Green Revolution
in cereal grain production (Delgado et al. 1999). From around
1980 to 2004, world meat production nearly doubled to 260 mil-
lion tons (FAO 2006a). To meet the growing demand, livestock
production may double again by 2020 (Pearson et al. 2005).
Animal agriculture worldwide is increasingly moving from
the relatively low efficiency, family-centered, low-input model
to intensive systems, which are loosely defined as the production
of large numbers of genotypically-similar animals often under
concentrated confinement with rapid population turnover. Tra-
ditional systems are being replaced by intensive systems at a
rate of more than 4% per year, particularly in Asia, Africa, and
South America (Pearson et al. 2005).
Intensive animal agriculture began in the United States with
poultry production more than 60 years ago and has become
the norm for both poultry and swine production (Pearson et al.
2005). The vast majority of chickens and turkeys raised for meat
in industrialized countries, and increasingly throughout the de-
veloping world (Mason and Crawford 1993), are now typically
reared in long sheds, each containing 15,000 to 50,000 birds
(Otte et al. 2007). Worldwide, an estimated 70-80% (UEP 2004)
of egg-laying chickens are intensively confined in battery cages,
small, barren, wire enclosures often stacked several tiers high
and extending down long rows in windowless sheds. It is not
uncommon for egg producers to keep hundreds of thousands
or even more than one million hens confined on a single farm
(Mench 2002). It is estimated that half of the world’s pig popu-
lation, which now approaches one billion animals, are raised in
industrial confinement (Nierenberg 2005).
While heralded for its efficiency and productive capacity, this
industrial model has raised sustainability concerns regarding the
waste absorption and feed supply capacity of available land (Del-
gado et al. 1999), as well as trepidation over associated zoonotic
risks (WHO 1999). This intensive industrialization of animal
agriculture may represent the most profound alteration of the
animal-human relationship since domestication (Delgado et al.
1999). Given the emergence of some of humanity’s most impor-
tant diseases in the Neolithic era, there is concern that new threats
may arise from the broadly significant changes currently taking
place in global livestock production (Slingenbergh et al. 2004).
Reviewing the 177 human pathogens classified as emerging
or reemerging, Woolhouse and Gowtage-Sequeria (2005) iden-
tified changes in land use or agriculture as the main driver of
their appearance. Unnaturally high concentrations of animals
confined indoors in a limited airspace and producing significant
quantities of waste may allow for the rapid selection, amplifi-
cation, and dissemination of zoonotic pathogens (Pearson et al.
2005). In the United States, the average numbers of animals
in each chicken, pig, and cattle operation approximately dou-
bled between 1978 and 1992 (Tilman et al. 2002). According
to the last U.S. Department of Agriculture census, 6% of the
nation’s hog farms produced three-quarters of the pigs and more
than 90% of the nation’s egg-laying hens were reared on 2%
of the country’s egg farms (USDA 2002). The crowding of in-
creasingly greater numbers of animals into increasingly smaller
spaces has been identified as a critical factor in the spread of
disease (Delgado et al. 2003).
Physiological stress has been associated with both con-
finement (Andr´e and Tuyttens 2005) and the various physical
modifications visited upon farm animals without anesthesia
(Puvadolpirod and Thaxton 2000), including castration, brand-
ing, dehorning, detoeing, teeth clipping, beak trimming, and
tail docking (Battaglia 2001). This, coupled with the metabolic
demands of intensive production, such as artificially augmented
reproduction, lactation, early weaning, and accelerated growth
rates, may leave animals extremely prone to disease (Cameron
2000).
The loss of genetic diversity in modern industrial livestock
production may also play a role in increased disease suscepti-
bility (USDA ARS 2006). Over the last century, an estimated
1,000 farm animal breeds have disappeared, and another 1,000
breeds are presently at risk for extinction—about one-sixth of
the world’s cattle and poultry varieties—in large part due to the
dilution or displacement of native breeds in developing countries
with breeding stock from industrialized countries (FAO 2004a).
This erosion of biodiversity may have public health impli-
cations. Genetic “bottlenecking” created by narrowly focused
breeding schemes may fuel zoonotic pathogen adaptation and
hinder host evolution for resistance (Ebert and Hamilton 1996).
Plant biologists have long understood the devastating disease po-
tential associated with genetic uniformity (Adams et al. 1971).
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
254
M. GREGER
In the early 1970s, for example, the U.S. corn industry developed
“Tcms” corn, a highly profitable strain adapted for large-scale
farming. Only after 85% of the nation’s seed corn acreage was
covered with the new variety did the industry realize that the
strain also happened to be particularly susceptible to a rare form
of leaf blight fungus that then wiped out large swaths of the
U.S. Corn Belt (Schrag and Wiener 1995). Emerging infections
may thrive when genetically homogenous members of a single
species are kept in close proximity, and the expanding applica-
tion of cloning technology may further narrow the genetic base
(Meredith 2004).
Genetic manipulation for production traits such as rapid
growth may itself carry a trade-off in immunocompetence
(Norris and Evans 2000). The “resource-allocation hypothesis”
suggests that protein and energy diversion from host defense to
breast muscle mass production in meat-type breeds of chick-
ens, for example, may explain why chickens with accelerated
growth are at risk for increased immune dysfunction (Gross and
Siegel 1988), disease morbidity (Rauw et al. 1998), and disease
mortality (Yunis et al. 2000). According to one economic anal-
ysis, transitioning to slower-growing breeds of broiler chickens
with improved immunity might be expected to cost poultry con-
sumers no more than a dollar or two a year (HSUS 2006). Given
the emergence of avian influenza viruses with putative pandemic
potential, this may be a small price to pay for a more disease
resistant global flock. Growth/productivity and disease suscep-
tibility have also been shown to be correlated in pigs (Hawken
et al. 1998), cattle (Mangel and Stamps 2001), and dairy cows
(Sinclair et al. 1999).
The amount of manure produced by high-throughput animal
husbandry creates a challenge to the maintenance of hygienic
standards. In the United States, farm animals produce an es-
timated 2 billion tons of manure each year (Loehr 1978), the
weight of 20,000 Nimitz-class aircraft carriers (NN 2006). Each
steer can produce 75 pounds of manure a day, potentially layer-
ing feedlots with waste (Purdue Research Foundation 2001). The
disease potential of high-density production has been compared
to that of cities in the Middle Ages where squalid overcrowding
facilitated the sewage-borne transmission of cholera and typhoid
before the Sanitary Revolution of the 19th century (Tauxe 2002).
Ironically, as the sanitary standards of human populations sub-
sequently rose, that of farm animal populations arguably may
have fallen (Muir and Aggrey 2003).
The industrialization of animal production may lead not only
to greater animal-to-animal contact, but increasing animal-to-
human contact, particularly when production facilities border
urban areas (Murphy 1999). Though land pressures have tended
to push crops and extensive animal systems away from the de-
veloping world’s growing megacities, intensive livestock oper-
ations are moving closer to major urban areas in countries such
as Bangladesh. This nexus, described as the “peri-urbanization”
of industrial animal agriculture (Slingenbergh et al. 2004), may
provide “flash points” for the source of new diseases (Delgado
et al. 2003).
A. Emerging Foodborne Zoonoses
More than five tons of animal manure are produced per human
capita in the United States every year (Loehr 1978), which may
affect human health directly or indirectly should it contaminate
food or water sources. An example of indirect harm is the emer-
gence of a novel toxicosis caused by Pfiesteria piscicida linked
to poultry manure runoff (Mahy and Brown 2000). This toxic di-
noflagellate has been implicated in the deaths of an estimated one
billion fish and skin lesions and neurological deficits reported in
boaters, swimmers, and Pfiesteria researchers (Burkholder and
Glasgow 1995).
A more direct harm can result when livestock manure contam-
inates municipal water supplies (Hrudey et al. 2003). In 1993,
cattle effluent upstream of a failing water treatment plant sup-
plying Milwaukee may have resulted in a record estimate of
403,000 cases of gastroenteritis caused by the zoonotic proto-
zoan Cryptosporidium (Mac Kenzie et al. 1994), although sub-
sequent evidence suggested a human rather than bovine fecal
source (Sulaiman et al. 1998). Manure-contaminated irrigation
water was presumed to be the source of the largest recorded out-
break of Escherichia coli O157:H7, affecting more than 7,000
Japanese schoolchildren who had eaten contaminated sprouts in
1996 (Michino et al. 1999).
Animal manure has been found to be the source of more than
100 zoonotic pathogens (Walton and White 1981) that may di-
rectly contaminate the food supply. Between the two decades
from 1975, when the dean of Yale’s School of Medicine fa-
mously told students that there were “no new diseases to be
discovered” and 1995, 17 new foodborne pathogens were iden-
tified (Liang 2002). Approximately half of all known foodborne
pathogens may have been discovered within the past 25 years
(Tauxe 2002).
In industrialized countries, the incidence of reported infec-
tious food- and waterborne illnesses has more than doubled
since the 1970s (Waltner-Toews and Lang 2000). An estimated
76 million Americans are stricken annually (Mead et al. 1999).
Worldwide, foodborne microbial disease carries an annual esti-
mated mortality of 20 million, with animal products topping the
list of causes (Delgado et al. 1999).
The global rise in incidence in foodborne diseases has been
attributed to both the greater consumption of animal-origin foods
and the methods of intensive production required to supply such
a demand (Adams and Motarjemi 1999). In the IOM report on
emerging infections, cattle feedlots and large-scale poultry pro-
duction were specifically singled out (Lederberg et al. 1992).
In the United States, the majority of the 9.3 billion chickens
raised each year (FAO 2007b) are typically stocked in densi-
ties between 14 and 23 birds per square meter (Fairchild 2005;
UEP 2006). In addition to higher potential zoonotic disease risk
in congested poultry operations (Shane 2000), higher stocking
densities in other species may be associated with elevated risk
of livestock infection with a number of potential zoonotic par-
asites and pathogens: Yersinia enterocolitica in goats (Lanada
et al. 2005); Trichostrongylus in sheep (Thamsborg et al. 1996);
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
255
M. bovis (White and Benhin 2004), Brucella (Salman and Meyer
1984), Salmonella (Jones et al. 1983), Neospora (Sanderson
et al. 2000), and Cryptosporidium in cattle (Atwill et al. 1999);
E. coli O157:H7 in both sheep and cattle (Stacey et al. 2007); Os-
tertagia in calves (Nansen et al. 1988); and Oesophagostomum
(Thamsborg et al. 1999), Aujeszky’s disease virus, and swine
flu virus in pigs (Maes et al. 2000).
The greatest stocking density in animal agriculture is likely
found in aquaculture, where tanks may hold up to one ton of
live weight in nine cubic meters of water (Kite-Powell 2004)
and floating cages have been known to hold up to four times this
density (Chapman 1992). The dense monoculture, poor water
quality, and resultant physiologic stress inherent to the “Blue
Revolution” rise in industrial aquaculture in the 1980s (FAO
2000) may have played a role in the emergence of the newly
discovered zoonotic pathogen Streptococcus iniae (Smith et al.
2001), which started devastating fish stocks on intensive fish
farms around the world with epidemics of acute meningitis (Goh
et al. 1998). The first case series of confirmed spread to humans
handling infected fish was reported in 1995: four cases of bac-
teremic cellulitis and one case of endocarditis, meningitis, and
probable septic arthritis (Weinstein et al. 1997). Wild fish mor-
talities observed proximate to mariculture S. iniae epidemics
suggest a risk to wildlife as well (Zlotkin et al. 1998). The rapid
expansion of intensive fish farming has also been associated with
an escalation in human Mycobacterium marinum infection and
disease (Lehane and Rawlin 2000).
1. Enterohemorrhagic E. coli
Humans play little role in the epidemiology of E. coli
O157:H7. It causes no signs of illness in its presumed natu-
ral host, cattle (Dean-Nystrom et al. 1998), but in humans, has
a relatively low infectious dose and may result in hemorrhagic
colitis and hemolytic uremic syndrome (Karmali et al. 1983).
As a commensal bovine bacterium, why did E. coli O157:H7
and the prophage it carries acquire the code for verotoxin pro-
duction, in part responsible for the morbidity and mortality of
colonized humans? Bovine intestinal cells have receptors for the
toxin (Menge et al. 2004), which in turn may promote intestinal
colonization (Robinson et al. 2006) via attaching and effacing
lesions (Sheng et al. 2006). This may be another example, as
perhaps is the case with SARS and civets, of pathogen exap-
tation, in which adaptation to a nonhuman host may increase
zoonotic potential.
E. coli O157:H7 was first described as an emerging foodborne
zoonotic pathogen in 1982, associated with consumption of un-
dercooked beef (Riley et al. 1983). The rise in E. coli O157:H7
as a human pathogen has been attributed to two primary fac-
tors: the demand-driven intensification of beef production and
the widespread use of antibiotics in animal agriculture. Indeed,
studies in Canada (Valcour et al. 2002), Sweden (Kistemann
et al. 2004), and France (Haus-Cheymol et al. 2005) have posi-
tively correlated regional cattle density with the rates of human
infection, and a number of antimicrobial agents have been shown
to enhance E. coli O157:H7 toxin production, as well as induce
the release of verotoxin-phages (Kohler et al. 2000), which may
have enhanced environmental stability compared to their bacte-
rial host (Moce-Llivina et al. 2003).
2. Listeria Monocytogenes
In 2002, E. coli O157:H7 resulted in one of the largest meat
recalls in U.S. history—19 million pounds of beef (USDA FSIS
2002a)—eclipsed only three months later with the recall of 27
million pounds of fresh and frozen poultry products contami-
nated with Listeria monocytogenes (USDA FSIS 2002b). Al-
though first identified in 1924 (Murrayet al. 1926), clinical Lis-
teria isolates were a laboratory rarity until the late 1970s when
reports increased in frequency and a series of human epidemic
outbreaks in the 1980s in North America (Fleming et al. 1985)
and Europe (Bille 1990) solidified its new status as an impor-
tant foodborne pathogen associated with ready-to-eat processed
meat and dairy products (Rocourt 1996).
L. monocytogenes has been described as having a “Jekyll
and Hyde” personality (Gray et al. 2006). In the natural envi-
ronment, it exists as a benign saprophyte living off decaying
vegetation (Fenlon 1999). In humans, it is one the deadliest bac-
terial infections currently known, with a mean mortality rate of
20–30% or higher despite early antibiotic treatment (Schuchat
et al. 1991). An invasive intracellular pathogen causing granulo-
matosis infantiseptica in term fetuses (Klatt et al. 1986), menin-
gitis in neonates (Synnott et al. 1994), and meningoencephalitis
in adults (Lorber 1996), L. monocytogenes may be responsible
for 28% of deaths due to foodborne illness in the United States
(Mead et al. 1999).
What is responsible for this transformation from an environ-
mental soil bacterium to public health hazard? When Listeria
cells are ingested by livestock exposed to spoiled feed (Wesley
1999), adaptation to their new environment commences with
the activation of a set of virulence-associated stress mediators
(Vazquez-Boland et al. 2001). This can lead to udder invasion
in mammals, resulting in mastitic milk (Bourry and Poutrel
1996), and fecal shedding for further transmission (Nightingale
et al. 2004). Though cattle tend to have a higher risk of ex-
posure via contaminated silage, poultry products are typically
more contaminated than beef, likely due to greater carcass cross-
contamination in poultry processing plants (Richmond 1990).
The environmental robustness inherent in soil microbes may
help L. monocytogenes survive food-processing technologies
such as salting, acidification, and refrigeration of contaminated
meat and dairy products (Lammerding and Doyle 1990). As
seen analogously with other zoonotic pathogens, there is evi-
dence that compared to Listeria strains found in the environment,
those adapted to growth in animals are more often associated
with large-scale human outbreaks (Graves et al. 2005).
3. ExPEC
Though E. coli O157:H7 remains the leading cause of acute
kidney failure in previously healthy U.S. children (Thorpe 2004),
fewer than 75,000 Americans become infected every year and
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
256
M. GREGER
fewer than 75 die, at a cost to the U.S. healthcare system of less
than 500 million USD (CDC 2005a). An estimated 150 million
Americans contract E. coli urinary tract infections (Stamm and
Norrby 2001), however, resulting in an estimated 36,000 dying
from sepsis at an annual cost in the billions (Russo and Johnson
2003).
Urinary tract infections (UTIs) are the most common bacterial
infections in women of all ages (Beers and Berkow 2005) and
among the most common bacterial infections around the world
overall (Foxman 2002). From a physician’s perspective, UTIs
are increasingly harder to treat as antibiotic resistance among the
chief pathogen, uropathogenic E. coli (UPEC), becomes more
common (CDC 2005c). It is generally agreed that the immediate
UPEC source causing human UTIs is an individual’s own colonic
flora (Johnson and Stell 2000), but mystery has surrounded how
these virulent clones come to inhabit the colon in the first place.
A recent hypothesis posits that most UTIs may be of zoonotic
origin (Rodriguez-Siek et al. 2005).
Pathogenic strains of E. coli have long been recognized
as sources of foodborne diarrhea (Smith et al. 2007). Strains
that cause disease outside the intestine are known as extrain-
testinal pathogenic E. coli (ExPEC) and include UTI-causing
uropathogenic E. coli (UPEC), newborn meningitis-causing
E. coli (NMEC), and avian pathogenic E. coli (APEC). The
zoonotic hypothesis suggests that retail poultry harboring APEC
represents a foodborne source of E. coli clones capable of caus-
ing human UTIs (UPEC) and possibly newborn meningitis as
well (Rodriguez-Siek et al. 2005).
A prospective two-year market survey of 1,648 diverse food
items from ten retail markets in the Minneapolis–St. Paul area
detected the presence of E. coli in 69% of the pork and beef
tested and 92% of the poultry samples, consistent with the con-
tamination of carcasses with fecal flora during slaughter and
processing. Half of the poultry samples were contaminated with
extraintestinal pathogenic E. coli bacteria, supporting the notion
that human UTI-type E. coli may be a foodborne pathogen. In-
terestingly, significantly less E. coli contamination was found
on meat samples purchased from natural food stores (Johnson
et al. 2005).
A number of studies have since indicated a commonality of
virulence factors in human and animal strains of ExPEC (Smith
et al. 2007). Moulin-Schouleur et al. (2007) used PCR-based
phylotyping and multilocus sequence typing to determine the
phylogenetic relationships between 39 APEC strains and 51
human ExPEC strains. The research team also compared the
virulence genotype and pathogenicity of APEC strains and hu-
man ExPEC in laboratory chicks. They concluded that APEC
strains belong to the same highly pathogenic clonal group as E.
coli strains of the same serotypes isolated from human cases of
neonatal meningitis, urinary tract infections, and septicemia.
Johnson et al. (2007) were the first to fully sequence an APEC
isolate and found compelling genomic similarity between it and
several human ExPEC strains. Based on a screening of a large
collection of APEC, UPEC, and NMEC strains for virulence
features, O-types, and phylogenetic background, Ewers et al.
(2007) speculated that not only may poultry be a vehicle or
even reservoir for human ExPEC strains, but may also serve
as an armory of virulence-associated genes that may transform
commensal human gut E coli into extraintestinal UPEC and
NMEC strains. Another recent report found that APEC plasmids
can contribute to the urovirulence of E. coli in mammalian hosts,
supporting this line of reasoning (Skyberg et al. 2006).
These studies bolster the hypothesis that at least some APEC
strains are a source of UPEC, indicating that these organisms po-
tentially represent a new class of foodborne zoonotic pathogens
(Smith et al. 2007). Validation of this hypothesis will require as-
sessment of the ability of APEC to persist on retail poultry, tra-
verse the human intestinal tract, and ascend and colonize the hu-
man urinary tract. Epidemiological studies will also be required
to confirm that such a phenomenon actually occurs (Johnson
et al. 2007). A broader geographic sampling of poultry products
would also be useful (Smith et al. 2007).
In chickens, APEC causes colibacillosis, now one of the most
significant and widespread infectious diseases in the poultry in-
dustry worldwide (Barnes et al. 2003). In a prospective study of
broiler chicken operations in Canada ranging in size from 7,700
to more than 100,000 birds, the risk of pathogenic E. coli infec-
tion was significantly associated with both increased farm size
and chicken stocking density (Elfadil et al. 1996). In caged egg-
laying hens, the most significant risk factor for flock infection
was hen density per cage. The less average space each bird had,
the higher the risk of outbreak. Vandekerchove et al. (2004) cal-
culated that every additional liter of cage living space afforded
to each hen would be associated with a corresponding 33% drop
in the risk of colibacillosis outbreak and suggested that new
European laws coming into effect implementing more spacious
cages would have a beneficial effect on the layers’ health. If
the zoonotic origin of UPEC is established, such laws may have
beneficial effects on consumers as well.
4. Multidrug-Resistant Bacteria
The emergence and dissemination of multidrug-resistant
zoonotic bacterial pathogens are increasingly being recognized
as a public health problem of global significance (Moore et al.
2006). The use of growth-promoting antibiotics in industrial
animal agriculture may be responsible for the majority of the in-
creases in antibiotic-resistant human isolates reported (Tollefson
et al. 1999).
The discovery that antibiotics could expedite growth in chick-
ens was announced at the annual meeting of the American
Chemical Society in 1950. By 1951, the U.S. Food and Drug
Administration (FDA) approved the addition of penicillin and
tetracycline to chicken feed as growth promoters, setting the
stage for pharmaceutical companies to mass-produce antibiotics
for animal agriculture (Boyd 2001).
The majority of present-day global antibiotic production is
used in animal agriculture (Tilman et al. 2002). Mellon et al.
(2001) estimate that up to 70% of antimicrobials used in the
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
257
United States are used as feed additives for chickens, pigs, and
cattle for non-therapeutic purposes. Additionally, three antibi-
otics have been approved for use in the U.S. aquaculture industry,
which is estimated to consume in excess of 20,000 kilograms of
antibiotics annually. Given the scale and intensification of the
broiler industry, though, chickens raised for meat have histori-
cally consumed the largest share (Viola and DeVincent 2006).
By the 1970s, nearly all commercially raised poultry in the
United States were reared on antibiotic-containing feed. By the
end of the 1990s, U.S. poultry producers may have been using
more than 4.5 million kilograms of antibiotics a year, a 300%
increase from the 1980s on a per-bird basis (Mellon et al. 2001).
The proportion of antibiotics used in livestock production for
growth promotion rather than therapeutic purposes may exceed
90% (Anderson et al. 2003).
Why the low-level feeding of antibiotics promotes faster
weight gain in animals raised for meat is not completely under-
stood (Graham et al. 2007). Germ-free chicks raised in germ-free
environments seem to grow faster than chickens in unsanitary en-
vironments, suggesting that immune activity may divert energy
from maximal growth (Klasing et al. 1987). Indeed, germ-free
chicks administered antibiotics experience no change in growth
rates, whereas commercially raised chickens demonstrate a re-
markable spurt in growth (Freeman et al. 1975). In the rela-
tively unsanitary conditions of intensive animal husbandry, nor-
mal physiological processes such as growth may be impaired in
light of the infectious load to which animals are exposed, which
may be mitigated by a constant influx of antibiotics (OTA 1979).
Although the use of antibiotics in livestock production can
lead to an improvement in feed efficiency and carcass quality
(Andreasen et al. 2005), the indiscriminate use of antibiotics
may select for drug-resistant zoonotic pathogens. Antibiotics
and antibiotic-resistant bacteria have been detected in the air,
groundwater, and soil around farms, and on retail meat (Smith
et al. 2005). In the Minneapolis-St. Paul ExPEC market survey,
for example, more than 80% of the E. coli recovered from beef,
pork, and poultry products were resistant to one or more an-
tibiotics, and greater than half of the isolates from poultry were
resistant to more than five drugs (Johnson et al. 2005). Humans
may then be exposed to these pathogens through infected meat,
vegetables fertilized with raw manure, or water supplies con-
taminated by farm animal waste (Acar and Moulin 2006).
At least 17 classes of antimicrobials are approved for farm
animal growth promotion in the United States, including β-
lactam, tetracycline, aminoglycoside, and macrolide classes,
all of which are important for clinical medicine (Anderson
et al. 2003). Indications of adverse human health implica-
tions have been derived from multiple lines of evidence: epi-
demiological studies tracing drug-resistant human infections to
specific farm animal production facilities, timelines showing
antibiotic-resistant infections in farm animal populations pre-
ceding the emergence of the same resistance in humans, and mi-
crobial studies showing that antibiotic-resistant zoonotic bacte-
ria may transfer that resistance to commensal human gut bacteria
(WHO/FAO/OIE 2003). Garofalo et al. (2007) recently demon-
strated that antibiotic-resistance genes could be detected directly
in chicken meat and pork. The strongest evidence may be data
from the European Union’s experience, which showed that after
antibiotics of human importance were banned for growth pro-
motion in 1998, there was a subsequent decrease in the levels of
antibiotic-resistant bacteria in farm animals, on meat, and within
the general human population (Smith et al. 2005). Recently a
large poultry corporation announced a reduction in antibiotic
use in feed (Tyson Foods 2007), and, in 2005, the quinolone
enrofloxacin was withdrawn from the U.S. market (FDA 2005).
4a. Campylobacter. Campylobacter species have long been
associated with livestock diseases, but only since the 1970s have
they been recognized as significant zoonotic poultry pathogens
(Moore and Matsuda 2002). Campylobacter spp. are currently
one of the leading causes of bacterial gastroenteritis in the world,
resulting in an estimated 1.4 million infections, 13,000 hospi-
talizations, and 100 deaths annually in the United States alone
(Friedman et al. 2000). The role of poultry in human Campy-
lobacteriosis was clearly demonstrated in Belgium in 1999 when
the discovery of high dioxin levels in feed led to a temporary
deficit in the retail sale of poultry and eggs. This resulted in
an estimated 40% reduction in the numbers of reported human
Campylobacter cases, which rose back to baseline when normal
poultry consumption levels resumed (Tenover et al. 2002). A
similar phenomenon was reportedly recorded in the Netherlands
in 2003 when avian influenza led to the depopulation of 30 mil-
lion birds (Rosenquist et al. 2004). Campylobacter colonization
of chickens is typically asymptomatic in the birds themselves
but can lead to significant fecal shedding (Doyle and Erickson
2006). A single, contaminated drop of raw retail chicken “juice”
may be enough of an infectious dose (McNamara 2002), and in-
fectivity may persist on a cutting board-type surface for hours
(Cools et al. 2005).
Typically, Campylobacteriosis is a self-limited, diarrheal
illness only requiring fluid replenishment, but antimicrobial
treatment may be prudent for patients with severe, prolonged,
or systemic infections. If bacteremia is suspected, a flu-
oroquinolone antibiotic such as ciprofloxacin is commonly
prescribed (Peterson 1994). Fluoroquinolone antibiotics have
been used in human medicine since 1986, but widespread
fluoroquinolone-resistant Campylobacter did not arise in the
United States until after quinolones such as enrofloxacin were
licensed for use in chickens in 1995 and 1996 for poultry house-
wide administration in drinking water (Gupta et al. 2004). In
countries such as Australia, which reserved fluoroquinolones
exclusively for human use, resistant Campylobacter have re-
mained rare (Price et al. 2005).
The FDA developed a quantitative risk assessment model that
suggested the consumption of an estimated 1.2 billion pounds of
boneless chicken contaminated with fluoroquinolone-resistant
Campylobacter in the United States in 1999 led to 153,580 hu-
man infections and a delay in effective treatment in 9,261 Amer-
icans (FDA 2001). Helms et al. (2005) showed that such a delay
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
258
M. GREGER
may lead to more complications and higher risk of adverse out-
come.
A lengthy legal process to stop the use of enrofloxacin in
poultry culminated in its withdrawal from the market, effective
September 2005 (Nelson et al. 2007). As the United States is
the world’s largest producer and exporter of poultry meat (FAO
2007b), this decision is expected to have a significant impact,
though fluoroquinolone-resistant Campylobacter spp. are ex-
pected to continue to circulate in poultry flocks for some time
(Humphrey et al. 2005). Meanwhile, U.S. Campylobacter iso-
late resistance to erythromycin continues to climb, and, in 2005,
the first multidrug-resistant isolate was detected, C. jejuni re-
sistant to ciprofloxacin, erythromycin, and ceftriaxone (Moore
et al. 2006).
4b. Salmonella. The vast majority of foodborne Campy-
lobacter infections involve an acute, self-limited, diarrheal
episode, but have been linked to the development of a rare in-
testinal lymphoma (Lecuit et al. 2004) and are now the suspected
cause of up to 25% of U.S. cases of irritable bowel syndrome
(Samuel et al. 2004). Approximately 1 of 1,000 Campylobacter
infections results in Guillain-Barr´e syndrome, a paralytic neu-
rological complication (Peterson 1994). With the virtual elim-
ination of polio in the developed world, this zoonotic poul-
try pathogen is now the most common cause of acute flaccid
paralysis in developed countries (Tauxe 2002). Similarly, up
to 15% of those suffering from Salmonella infection may de-
velop a chronic condition of joint inflammation. An estimated
100,000 to 200,000 Americans suffer from arthritis arising di-
rectly from foodborne infections every year (McDowell and
McElvaine 1997).
Annually in the United States, nontyphoidal Salmonella in-
fections kill hundreds, hospitalize thousands, and may sicken
more than a million (Mead et al. 1999). Salmonella infections are
often also self-limited, but may progress to bacteremia in 3–10%
of culture-confirmed cases, particularly in infants (Stutman
1994), for which antimicrobial therapy can be life saving (Tauxe
1991). In adults, fluoroquinolones such as ciprofloxacin are the
antibiotic of choice, while extended-spectrum cephalosporins
such as ceftriaxone are considered preferred treatments for chil-
dren (Fey et al. 2000). Growing Salmonella resistance to both
frontline therapies has elicited global concern.
Multidrug-resistant Salmonella enterica serovar Ty-
phimurium phage type 104 (DT104), which is resistant to up
to six common antimicrobials including ciprofloxacin, rapidly
emerged in the 1980s as a global health problem. DT104
was first detected in humans and cattle but has since become
common in poultry, particularly turkeys, as well as pigs and
sheep (Threlfall 2000). Contact with sick farm animals and the
consumption of chicken, pork sausages, and meat paste were
early identified risk factors for DT104 infection, elevating the
urgency of debate over the clinical consequences of drug resis-
tance in zoonotic bacteria (Wall et al. 1994). Once acquired, the
cluster of genes coding for multidrug resistance may be passed
horizontally to other serovars, as may have been documented
in poultry in Belgium (Cloeckaert et al. 2000). Comparing
outbreaks caused by resistant versus pansusceptible Salmonella
isolates, elevated morbidity (hospitalization) and mortality rates
illustrate the clinical consequences of antimicrobial resistance
(Holmberg et al. 1984).
The emergence and worldwide dissemination of DT104 has
been speculatively blamed on antimicrobial use in aquaculture.
DT104 chloramphenicol resistance is due to floR, the florfenicol-
resistance gene first identified in a piscine bacterium (Bolton
et al. 1999) and presumably selected for by the use of florfenicol
in aquaculture operations in Asia (Sano 1998). DT104 tetracy-
cline resistance is likewise caused by a gene first identified in a
fish pathogen (Zhao and Aoki 1992), and the DNA sequence en-
compassing both resistance genes on the DT104 chromosome is
closely related to a plasmid first isolated in a third bacterial fish
species in aquaculture operations in Japan (Kim and Aoki 1993).
Furushita et al. (2003) showed that tetracycline-resistance genes
found in hatchery tanks at fish farming operations and clinical
(hospital effluent) bacterial isolates shared 92–100% sequence
identity, and that resistance was experimentally transferable,
suggesting that the human (hospital) and aquaculture compart-
ments of the global environment behave interactively. The re-
sistance determinants of DT104 may therefore have emerged
among aquaculture bacteria and been horizontally transferred to
DT104 (Angulo and Griffin 2000).
Although farmed fish may serve as a vehicle for the transmis-
sion of resistance-coding plasmids to commensal or pathogenic
human bacteria (Rhodes et al. 2000), human-to-human transmis-
sion is so inefficient in industrialized countries (Cohen and Tauxe
1986) that the rapid global dissemination of these resistance
gene clusters in the 1980s must be explained by other means.
In the 1980s in Japan, where the use of florfenicol, tetracy-
clines, and 20 other antimicrobials are permitted in aquaculture
production (Sano 1998), exports of fish meal rose over 500%,
reaching 223,859 tons in 1989 (FAO 2007b). Combined with
the model presented by the international spread of a previous
Salmonella serotype traced to the global distribution of con-
taminated fish meal with subsequent livestock and human iso-
late recovery (Clark et al. 1973), DT104 resistance gene clusters
may have been distributed worldwide in the 1980s via contami-
nated livestock feed made out of farmed fish (Angulo and Griffin
2000). Perhaps because of host differences in thermoregulation
(Alderman and Hastings 1998), aquatic zoonotics may be rela-
tively rare, but antimicrobial resistance acquired by farmed fish
pathogens may horizontally transfer to pathogens of commonly
eaten homeotherms (Kruse and Sorum 1994).
Once regional distribution is achieved, the persistence and
spread of resistance gene clusters could be facilitated by the re-
cycling of livestock waste into feed (Turnbull 1979) and further
antimicrobial use. Preweaned U.S. dairy calves, for example,
are suckled with tetracycline-containing milk replacer, which is
commonly used commercially to prevent diarrhea thought exac-
erbated by maternal colostrum deprivation (Berge et al. 2005).
Berge et al. (2006) conducted a field trial that demonstrated
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
259
that the use of neomycin/tetracycline-medicated milk replacer in
preweaned calves selected for bacteria with resistance clusters
to antimicrobials not used in the study or on the ranch.
Concern over the promiscuity of self-transmissible plasmids
coding multidrug-resistance reached new levels with the detec-
tion of a multidrug-resistant strain of Yersinia pestis (IP275) in
1995 (Galimand et al. 1997). The etiological agent of plague,
Y. pestis,isazoonotic bacterial pathogen that has caused
multiple pandemics resulting in an estimated 200 million hu-
man deaths but, with antibiotic treatment and prophylaxis, has
been effectively controlled throughout most of the world (Perry
et al. 1997). Y. pestis IP275 contains a self-transmissible plas-
mid conferring high-level resistance to at least eight antimi-
crobials, including streptomycin, tetracycline, chloramphenicol
and sulfonamides, drugs that are recommended for use in plague
(Inglesby et al. 2000). Comparative DNA analysis of the Y. pestis
IP275 plasmid showed a near identical backbone to multidrug-
resistance plasmids isolated from a Salmonella spp. and a fish
pathogen, Yersinia ruckeri. Furthermore, Y. pestis IP275-like
plasmids were isolated from beef, chicken, turkey, and pork
meat samples over a wide geographic distribution, suggesting
that the agricultural use of antibiotics may have broader biomed-
ical, public health, and biodefense implications than previously
recognized (Welch et al. 2007).
The most dramatic emergence of multidrug-resistant
Salmonella over the last two decades may have been the ap-
pearance and global spread of Salmonella resistant to extended-
spectrum cephalosporins such as ceftriaxone, the third-
generation cephalosporin used to treat invasive blood infec-
tions in children (Arlet et al. 2006). In addition to its resistance
to five drugs typically found in DT104—ampicillin, chloram-
phenicol, streptomycin, sulfamethoxazole, and tetracycline—
Salmonella enterica serotype Newport isolates resistant to at
least nine antimicrobials (S. Newport MDR-AmpC) have arisen,
resistant to amoxicillin/clavulanate, cefoxitin, ceftiofur, and
cephalothin as well, with decreased susceptibility to ceftriax-
one (MIC ≥ 16 µg/ml) (Harbottle et al. 2006).
Salmonella serotype Newport has emerged as the third most
common serotype causing human salmonellosis in the United
States, with a doubling of laboratory-confirmed cases between
1997 and 2001. Human outbreaks have been linked to the con-
sumption of hamburger (Spika et al. 1987), pork (Narain and
Lofgren 1989), chicken (Anand et al. 1980), and fish (Heinitz
et al. 2000). The increase in S. Newport isolates is likely associ-
ated with the rapid dissemination of multidrug-resistant strains
(CDC 2002). By 2004, Salmonella resistance to extended-
spectrum cephalosporins was identified in 43 countries (Arlet
et al. 2006).
The first documented domestically-acquired U.S. case of
Salmonella infection resistant to ceftriaxone (along with 12
other antimicrobials) was isolated in 1998 and found closely
matched to isolates in a local outbreak of bovine salmonellosis,
providing additional evidence that antibiotic-resistant strains of
Salmonella evolve primarily in livestock (Fey et al. 2000). In
1998, 2% of Salmonella isolates collected from chickens and
cattle in sampled U.S. slaughter plants were resistant to ceftri-
axone (FDA/USDA/CDC 1998).
The use of ceftiofur in farm animals has come under in-
creasing scrutiny as a selective factor responsible for the emer-
gence and persistence of ceftriaxone-resistant enteric pathogens
such as S. Newport MDR-AmpC. Ceftiofur was the only
cephalosporin approved for systemic use in pigs, sheep, chick-
ens, and turkeys in the United States, and cross-resistance be-
tween ceftiofur and ceftriaxone (and cephamycins) has been
demonstrated. In this way, the widespread use of ceftiofur in ani-
mals raised for human consumption may have contributed to cef-
triaxone resistance in Salmonella, which may then be transmitted
to consumers through the food supply (Dunne et al. 2000). Con-
troversially, a fourth-generation cephalosporin, cefquinome, is
currently under review for FDA approval for use in cattle
(Feldgarden 2007).
Animals bearing Salmonella and Campylobacter can remain
healthy carriers with horizontal transmission mediated by fecal
shedding. Although Campylobacter colonization of the avian
reproductive tract may lead to external egg contamination (Byrd
et al. 2007), Salmonella can lead to ovarian infection, resulting
in vertical transmission and the potential for additional zoonotic
risk (Gast and Beard 1990).
The leading cause of salmonellosis in the world is now the
egg-borne serovar Salmonella enterica enteritidis (S. enteritidis)
(Baumler et al. 2000). With beginnings traced to the mid-1970s,
by the 1990s human S. enteritidis infections had reached pan-
demic proportions (Pang et al. 1995). In the United States, S.
enteritidis cases began to appear in the Northeast in the late
1970s and spread to the Mid-Atlantic States by the mid-1980s
(Braden 2006). A single U.S. outbreak in 1994, in which tanker
trailers previously carrying raw eggs contaminated the premix
of a nationally distributed ice cream, resulted in an estimated
224,000 human cases of S. enteritidis infection (Hennessy et al.
1996). On average, though, the CDC estimates that 50,000 to
110,000 S. enteritidis infections each year in the United States
can be attributed to eggs (Braden 2006).
In S. enteritidis outbreaks in which a food source was con-
firmed, 75% were traced to eggs or egg-containing products
(Braden 2006). Eggs are one of the few animal products that
are frequently eaten raw or undercooked, as in such items as
homemade or restaurant-produced salad dressings, hollandaise
sauce, mayonnaise, ice creams, or beverages including eggnog.
Humphrey et al. (1989) showed that “sunny-side up” frying, for
example, was insufficient to destroy S. enteritidis inoculated into
yolks at concentrations typical of transovarian infection.
S. enteritidis infections have fallen over the last decade in the
United States due largely to farm-based assurance programs,
early and sustained egg refrigeration, and educational efforts
targeted at food service workers and consumers (Braden 2006),
but these domestic efforts and successes may pale in comparison
to those made in other countries (Hopp et al. 1999). Though the
poultry industry operates under severe cost pressures, changes
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
260
M. GREGER
in management practices may be necessary to make significant
further progress in reducing this zoonotic risk (Humphrey 2006).
Holt et al. (1998), for example, showed that poor ventilation
and high dust levels appear to aid dissemination of S. enteritidis
among chickens via colonization of exposed mucosal surfaces,
suggesting that better poultry house design to improve air qual-
ity may decrease S. enteritidis infection rates. The high den-
sity of modern commercial chicken populations may not only
increase risk due to the number of potentially infectious con-
tacts and the level of airborne particulates and litter contami-
nation (Braden 2006), but may have physiological effects that
facilitate infection mediated through the sympathetic-adrenal
medullary axis. A range of food production-relevant stresses
have been shown to facilitate extra-intestinal invasion by bacte-
rial pathogens (Swildens et al. 2004).
Stress caused by severe overcrowding, for example, such
as in high-density layer battery caging (Humphrey 2006), has
been shown to increase Salmonella infection susceptibility in
mice (Kuriyama et al. 1996). The physiologic stress associated
with the forced molting of commercial egg-laying flocks to ex-
tend their productive life increases S. enteritidis shedding (Holt
et al. 1994) and may double the incidence of egg contamination
(USDA 1998). Noradrenaline (NA) is a catecholamine released
by animals under stress. Chickens who are acutely (Knowles
and Broom 1993) and chronically (Cheng et al. 2002) stressed
may have higher intestinal and circulating levels of NA. Nora-
drenaline is taken up by E. coli, increases bacterial growth rates
(Kinney et al. 2000), and may up-regulate certain virulence fac-
tors (Lyte 2004). In iron-restricted media used to simulate the in-
testinal mucosal surface, NA markedly increases Campylobacter
growth rates and may increase bacterial motility, colonization,
and the expression of virulence factors (Humphrey 2006). This
may contribute to the phenomenon of C. jenuni isolation from
deep muscle tissues in retail poultry (Humphrey 1991), which
may present a greater public health risk than surface contami-
nation. Stress-induced neurotransmitters may also increase the
growth rates of Salmonella spp. (Bailey et al. 1999), Listeria,
(Coulanges et al. 1998), and other zoonotic pathogens (Belay
et al. 2003), as well as increase the expression of virulence fac-
tors in Pseudomonas aeruginosa (Kinney et al. 2000).
5. Bovine Spongiform Encephalopathy
Certain modern agricultural practices to enhance productiv-
ity may have unintended public health consequences (Phua and
Lee 2005). The use of antimicrobials for growth promotion is
one such practice, and the continued feeding of slaughterhouse
waste, blood products, and excrement to livestock may be an-
other (Sapkota et al. 2007).
Annually, the United States produces the majority (NRA
2005) of the estimated 10 million tons of animal protein concen-
trates incorporated worldwide into livestock feed, such as meat,
blood, and bone meal (WHO and OIE 1999). Incomplete inacti-
vation of pathogens during processing may pose a risk of disease
transmission. U.S. livestock feeds incorporating meat byprod-
ucts in particular (McChesney et al. 1995) have been found to
be contaminated with a number of zoonotic pathogens such as
Salmonella spp. (Crump et al. 2002) and E. coli O157:H7 (Dar-
gatz et al. 2005). Testing 165 rendered animal protein products
originating from poultry, cattle, and fish at a poultry feed mill,
Hofacre et al. (2001) found antibiotic-resistant bacteria in 85%
of samples.
Bovine spongiform encephalopathy (BSE) may have arisen
because of the inclusion of the rendered remains of sheep in-
fected with an ovine spongiform encephalopathy in cattle feed,
suggesting that animal-based feed additives may play a role in
zoonotic disease emergence as well (Wilesmith et al. 1991). The
further recycling of the remains of infected cattle into cattle feed
likely contributed to the spread of the epizootic in the United
Kingdom, where BSE was first recognized in 1985 (Anderson
et al. 1996), and its subsequent spread to 24 countries (USDA
APHIS 2005a).
BSE belongs to a class of transmissible spongiform en-
cephalopathies thought caused by prions, an unconventional
class of proteinaceous pathogens uniquely resilient to food
preparation safeguards, demonstrating limited survival even af-
ter incineration at 360
◦
C. Exceptional prion resistance to heat,
formalin (Pattison 1965), and radiation (Gibbs et al. 1978) pre-
sumably derive from the putative lack of associated nucleic acid
(Prusiner 1998).
A British Medical Journal editorial described BSE as re-
sulting from an “accidental experiment on the dietary transmis-
sibility of prion disease between sheep and cows” (Harrison
and Roberts 1992). A subsequent “accidental experiment”—
with humans—may have started in the late 1980s, as tissues
from infected cattle increasingly entered the human food sup-
ply (Phillips et al. 2000). Approximately 100 human cases of
variant Creutzfeldt-Jakob disease (vCJD), the invariably fatal
neurodegenerative disease (Collinge 1999) thought caused by
the consumption of contaminated bovine products (Bruce et al.
1997), have been neuropathologically confirmed in the United
Kingdom (NN 2007b). The potentially decades-long incubation
period (Brown et al. 2000), combined with the discovery of an
infection-positive human appendectomy tissue sample removed
during routine surgery (Hilton et al. 2002), have raised concerns
about the possibility of secondary iatrogenic transmission via
blood products or surgical instruments (Zobeley et al. 1999). The
fourth case of probable transfusion transmission of vCJD infec-
tion in the United Kingdom was recently reported (NN 2007a).
Experimental verification of blood infectivity (Houston et al.
2000) has renewed criticism over the practice of weaning U.S.
dairy calves on milk replacer containing bovine blood products
(DG-SANCO 2002), given the presence of BSE in North Amer-
ica (USDA 2006). In 1997, the U.S. FDA restricted the use of
tissues considered particularly risky—bovine skull, brain, eyes,
parts of the vertebral column, spinal cord, trigeminal and dor-
sal root ganglia, tonsils, and distal ileum—in the formulation
of cattle and other ruminant feed, but continued to allow the
intraspecies recycling of ruminant blood (GAO 2002). Most
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
261
newborn calves in the United States are separated from their
dams within 12 hours, many immediately after birth, so the milk
can be marketed for human consumption. Though some dairy
producers still wean calves on whole milk, the majority report-
edly use milk replacer, which may contain spray-dried cattle
blood as a cheap source of protein (USDA APHIS 1993).
The bovine neurological and lymphoid tissues specified as
higher risk may not be fed legally to ruminants but are still
allowed to be incorporated into nonruminant feeds and sub-
sequent waste products are then permitted in ruminant feeds
(USDA FSIS 2005). The use of ruminant meat and bone meal
in the diets of broiler chickens, for example, may pose the risk
of cycling prions back to cattle if the broiler litter, which may
contain spilled feed and poultry waste, is then fed back to cattle
(Rankins et al. 2002). Although excrement from other species
is fed to cattle as well (Haapapuro et al. 1997), poultry waste is
considered more nutritious for cows than swine waste or cattle
dung (Fontenot et al. 1971). As much as eight times cheaper
than foodstuffs like alfalfa (Fontenot 2001), poultry litter has
been used as an economical feedstuff in the United States in
place of hay (Rude and Rankins 1997) since the 1950s (Noland
et al. 1955).
Two human outbreaks of Salmonella infection tied to the use
of meat and bone meal in feed were recorded in the United
Kingdom decades before the emergence of BSE (Knox et al.
1963; Pennington et al. 1968). Although the U.K. rendering in-
dustry may not have been able to foresee the full extent of the
potential hazard of continuing this practice, the U.S. render-
ing industry has no such justification. In general, the livestock
industries have reportedly tended to favor stop-gap measures
such as food irradiation rather than fundamental changes in farm
management practices (Nestle 2003). Investments in chemical
carcass dehairing technologies to lessen fecal contamination in
cattle processing plants (Nou et al. 2003) and the investigation of
bacteriophage application to fresh poultry (Doyle and Erickson
2006) are examples of this trend.
One reason poultry products may be such common carriers
of foodborne diseases (Gregory et al. 1997) is that unlike the
potentially manure-encrusted hides of cattle, for example, the
skin can be eaten with the meat. Efforts to improve on-farm hy-
giene to decrease eventual carcass surface contamination may be
more effective in the long run than recently proposed technolo-
gies such as employing equipment designed to exert abdominal
pressure to force out cloacal contents before scalding (Wagenaar
et al. 2006) and the proposed use of either mechanical plug, sta-
pling, or cyanoacrylate adhesive technology to seal the rectal
cavity of farmed birds during processing (NN 1995).
The BSE crisis in the United Kingdom affected major in-
stitutional reform in government infrastructure, leading to the
establishment of the Food Standards Agency independent from
the perceived bias towards producer interests in the Ministry of
Agriculture, Fisheries, and Food (Wales et al. 2006). The U.S.
congressional Government Accountability Office (GAO 2001)
and Institute of Medicine (IOM 1988) have called for the sim-
ilar formation of a U.S. food safety authority independent of
the USDA in hopes of further reducing the risk of foodborne
zoonotic infections.
B. Streptococcus Suis
China is the world’s largest producer of pork (FAO 2007b),
scaling up the size of commercial operations (Bean and Jianping
2005) and doubling production to more than 50 million tons
between 1990 and 2005, the year the largest pork-producing
province (Xin et al. 2001) suffered an unprecedented human
outbreak of the emerging zoonotic pig pathogen Streptococcus
suis (Lun et al. 2007). S. suis is a common cause of meningi-
tis in intensively farmed pigs worldwide (Sanford and Higgins
1992) and, similarly, presents most often as purulent meningitis
in humans who handle or butcher sick pigs (Huang et al. 2005).
Thought due to inner ear involvement (Kay 1991), 39% of hu-
man survivors may suffer permanent sensorineural hearing loss
(Wang et al. 2007).
In summer 2005 in Sichuan, China, a total of 215 probable
cases of human S. suis infection were reported—more than the
total number of previously recorded human cases worldwide
(Tang et al. 2006)—following a local epizootic involving ap-
proximately 80,000 pigs (Lun et al. 2007). Fatal outcomes were
reported in 39 (18%) of human cases (Yu et al. 2006), perhaps
triple the case fatality rate seen previously in sporadic human
cases elsewhere (Normile 2005). Though the increased mortal-
ity in China may reflect regional differences in treatment access
(Sriskandan and Slater 2006), most deaths appeared due to a
potentially novel form of invasive toxic shock syndrome, com-
bining deep-tissue infection and vascular collapse (Tang et al.
2006). In cases of S. suis septic shock, mortality may exceed
70% irrespective of treatment (Lun et al. 2007).
The first human case of S. suis infection was reported in
Denmark in 1968 (Arends and Zanen 1988). Cases have since
been reported in many countries with intensive swine produc-
tion (Lun et al. 2007) in which S. suis has emerged, particularly
during the past 15 years (Gottschalk and Segura 2000), as a
major cause of porcine infection and economic loss (Escudero
et al. 2007). In the United States, the world’s second-largest
pork producer (FAO 2007b), S. suis may cost the swine indus-
try more than 300 million USD annually (Staats et al. 1997).
Though human cases of S. suis meningitis may likely be under-
diagnosed and misidentified (Gottschalk 2004) due to the lack
of adequate surveillance (Cole et al. 2000), the first U.S. case
was reported in 2006 (Willenburg et al. 2006). The American
Association of Swine Veterinarians implicates increasing swine
industry intensification as a critical factor in the emergence of
porcine zoonoses in general, via both elevated environmental
pathogen loads and the impairment of porcine host immunity
(Meredith 2004). S. suis may fit this pattern (USDA APHIS
2005b).
S. suis is enzootic in most countries with intensively managed
pig populations with high rates of asymptomatic carriers docu-
mented. Clinically normal pigs may harbor S. suis in their nasal
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
262
M. GREGER
cavity, palantine tonsils, and genital and alimentary tracts, and
play an important role in the propagation of S. suis via “nose-
to-nose” oronasal transmission (Arends et al. 1984). Isolating
S. suis from tonsillar crypts at slaughter, Clifton-Hadley (1986)
reported herd carrier rates as high as 100%.
S. suis is likely transferred from vaginal secretions to the oral
cavity of the piglet during parturition and colonizes the ton-
sils soon after birth (Amass et al. 1996). While experimental
intravenous inoculation results in high morbidity and mortality
(Busque et al. 1997), intranasal inoculation, the presumed nat-
ural route, often results in no or limited disease, lesions, and
mortality (Iglesias and Trujano 1992). S. suis can become in-
vasive, though, and result in a wide range of diseases in pigs
(Staats et al. 1997). Pneumonia has historically been the primary
porcine presentation, but meningitis and endocarditis have be-
come more common, attributed to changing housing conditions
and management practices (Aarestrup et al. 1998).
The factors responsible for S. suis traversing the mucosal
barrier to induce disease are not known at this time (Gottschalk
and Segura 2000), but Pallares et al. (2003) showed that pre-
irritation of the nasal cavity with 1% acetic acid an hour be-
fore intranasal inoculation could result in invasive meningitic
disease. This suggested that mucosal damage, stress, or both,
induced by the acid pretreatment, allows S. suis to become sep-
ticemic. Once invasive, S. suis may then be spread via respiratory
droplets (Berthelot-Herault et al. 2001) or more directly via con-
tact with contaminated blood on improperly sterilized castration
scalpels, tooth-cutting pliers, or tail-docking knives (Du 2005).
At 25
◦
C, S. suis has been shown to survive for 24 hours in dust
and for eight days in feces (Clifton-Hadley and Enright 1984).
A number of factors may have contributed to the scale of
the 2005 outbreak in China, i.e., a (still putative) highly vir-
ulent S. suis clone, the eating of undercooked pork from sick
pigs, and delayed diagnosis and treatment (Lun et al. 2007). Fo-
cus, however, has fallen on the way in which pigs were raised.
Poor housing conditions, such as overcrowding, inadequate ven-
tilation, and related physiological stress, were considered risk
factors for the initial epizootic outbreak (WHO 2005).
China now has swine facilities confining more than 50,000
head (Simpson et al. 1999), which may have stocking densities
and sanitation levels exceptionally conducive to the emergence
and spread of this disease (Arends et al. 1984). Dee et al. (1993),
for example, showed experimentally that overcrowding and so-
cial stress led to higher-than-average S. suis infection rates. Co-
infection with porcine reproductive and respiratory syndrome
virus (PRRSV) or Aujeszky’s disease virus can exacerbate S.
suis infection in pigs (Iglesias et al. 1992). PRRSV is typically
enzootic in intensive pig production settings (Chung et al. 1997),
and Aujeszky’s disease, also on intensive pig farms (Chew-Lim
and Ng 1987), shows a stress-induced susceptibility in pigs (de
Groot et al. 2001) (but not mice) (de Groot et al. 1999), sug-
gesting that the association between S. suis infection and high-
density pig herds is likely multifactorial (Clifton-Hadley and
Enright 1984).
A similar pattern is found in other porcine respiratory dis-
eases, which have been linked to the increased crowding of pigs,
both per pen and per building (Enøe et al. 2002). Based on the
isolation of pathogens from air samples taken inside produc-
tion facilities, Madec and Rose (2003) suggested diminished air
volume per animal may increase the concentration of infectious
particles and thereby facilitate aerosol spread. This seems to be
the case with swine influenza risk, which has been associated
in commercial pig operations with increased density of pigs per
pen, pigs per operation, and pigs per municipality (Maes et al.
2000), though this may derive from a combination of pathogen
density and confinement-related stressors.
Breeding sows restricted to narrow stalls, as is common dur-
ing gestation and farrowing in intensive production, produce
lower levels of antibodies in response to an experimental chal-
lenge (Siegel 1983). Measures as simple as providing straw bed-
ding for pigs may improve immune function by eliminating the
immunosuppressive stress of lying on bare concrete (Andr´e and
Tuyttens 2005). Ewald et al. (1994), for example, found that
straw bedding was linked to decreased risk of swine flu infection.
Although S. suis was isolated from 6% of raw pork prod-
ucts sampled in Hong Kong in 2005 (Ip et al. 2007), data from
reportedly all epidemiological surveys on outbreaks in China
strongly indicate that direct contact with sick pigs is necessary
for zoonotic infection (Lun et al. 2007). The infection rate of pig
farmers, abattoir workers, meat-processing workers, and veteri-
narians has been estimated to be 1,500-times higher than that
of the general population (Arends and Zanen 1988). S. suis,
therefore, remains essentially an occupational zoonosis. Espe-
cially given concern over rising emergence of fluoroquinolone-
resistant strains now reported among clinical swine isolates (Es-
cudero et al. 2007), the exclusion of splenectomized individ-
uals from the pork trade has been suggested (Watkins et al.
2001).
C. Avian and Pandemic Influenza
The dozens of emerging zoonotic threats that have charac-
terized this third era of human disease must be put into context.
The total number of human cases reported of S. suis infection
worldwide is approximately 400 (Tang et al. 2006). Nipah virus
and BSE prions seem to have so far infected fewer still. SARS-
CoV did infect thousands and HIV millions, but there may be
only one known pathogen capable of rapidly infecting billions:
influenzavirus A.
Influenza has been called the “last great plague of man”
(Kaplan and Webster 1977). Unlike infectious diseases like
malaria, for example, which tend to be confined equatorially,
or HIV/AIDS, which is only fluid-borne, influenza may be the
only pathogen carrying the potential to infect a substantial per-
centage of the world’s population within a matter of months.
In the 4,500 years influenza virus has been presumed to infect
humans since the first domestication of wild birds (Shortridge
2003b), influenza has always been one of humanity’s most con-
tagious diseases (Mills et al. 2004). But only since 1997 with the
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
263
recognition of the zoonotic potential of highly pathogenic avian
influenzavirus A subtype H5N1 has influenza also emerged as
potentially one of the deadliest, with a reported case fatality
rate (CFR) exceeding 60% (WHO 2007). Although H5N1 has
resulted in more than 200 million birds killed or culled (FAO
2006b), the ratio of avian to human deaths exceeds a million
to one, with only approximately 200 confirmed human deaths
(WHO 2007).
Though infinitesimal compared to the 1.7 million human
deaths attributed to TB every year (WHO 2006a) or the more
than 3 million to AIDS (UNAIDS and WHO 2005), one rea-
son the emergence of H5N1 has raised so much concern is that
the last known time a wholly avian influenza virus resulted in
human mortality, it triggered the single deadliest disease event
in human history: the 1918–1919 influenza pandemic in which
30% of the world’s population may have fallen ill (Burnet and
Clark 1942) and 50 million may have died (Johnson and Mueller
2002).
H5N1 is now almost exclusively a disease of birds (Yang et al.
2007). Should the virus have the dual capacity to acquire efficient
human transmissibility while maintaining its human virulence,
H5N1 (CFR ∼60%) could theoretically trigger a pandemic that
makes the 1918 pandemic (CFR ∼2.5%) (Marks and Beatty
1976) seem in comparison like the seasonal flu (CFR ∼0.1%)
(Rosenau and Last 1980). This makes the factors underlying the
continuous emergence of this zoonotic virus particularly impor-
tant to understand.
1. Avian Influenza
1a. Shifting Ecology. Though first clearly described more
than 800 years ago (reviewed in Potter 2001), influenza can be
considered a perpetually emerging disease given its ability for
antigenic transformation. Type A viruses are subdivided into
subtypes according to the antigenic nature of two surface pro-
teins, hemagglutinin (HA) and neuraminidase (NA), of which
there are 16 (H1 to H16) and 9 (N1 to N9) known, respec-
tively (Fouchier et al. 2005). “H5N1” denotes that the virus en-
velope displays the fifth hemagglutinin type in the WHO-naming
scheme, along with the first neuraminidase type (WHO 1980a).
Laver and Webster (1972) were the first to suggest through
ecological studies that aquatic birds are the natural reservoir
for the virus, which, in wild ducks, tends to exist as an asymp-
tomatic gastrointestinal infection (Webster et al. 1978). More
recent recognition has been made of the possibility that shore-
birds may also maintain part of the influenza gene pool (Sharp
et al. 1993). In nature, then, the influenza virus has likely existed
for millions of years as a harmless (Kida et al. 1980), enteric,
waterborne infection of waterfowl (Markwell and Shortridge
1982).
Studies of ducks clustered on Canadian lakes show that up
to 30% of juvenile birds may be actively shedding the virus
(Hinshaw et al. 1980b). Most infected ducks evidently only shed
virus for a few days (Webster et al. 1993), but the presumed
fecal-water-oral route of infection for aquatic birds (Markwell
and Shortridge 1982) is thought efficient enough to have kept
the virus circulating throughout the millennia (Shortridge 1992).
The ducks excrete such high titers of virus that Hinshaw et al.
(1979) were able to culture influenza viruses straight from un-
concentrated lake water. Under permissive conditions, the virus
is estimated to be able to persist for years in cold water (Wobeser
1997) and possibly even survive in ice (Zhang et al. 2006). Given
such high concentrations of virus, the efficiency of transmis-
sion (Webster et al. 1997), and with such environmental sta-
bility, influenzavirus A likely infects virtually all of the mil-
lions of ducks in the world sometime within their lives (Murphy
1993).
Human influenza pandemics are thought triggered by genetic
infusion from this avian reservoir, resulting in influenzavirus A
subtypes with surface proteins to which the present human gen-
eration is immunologically na¨ıve (Webster and Laver 1971). A
purely avian H1N1 subtype is believed to have caused the 1918-
1919 pandemic, presumably via adaptive mutation (Tauben-
berger et al. 2005), infecting an estimated one billion people over
12 months (Oxford et al. 2006). The “Asian flu” pandemic of
1957 and the 1968–1969 “Hong Kong flu” were generated when
the circulating seasonal human subtype acquired new avian gene
segments via reassortment (Scholtissek et al. 1978), each result-
ing in approximately one million deaths (Pyle 1986). One specu-
lation as to why the 1918 pandemic virus may have cost as many
as 25 times more human lives than both these pandemics com-
bined (Johnson and Mueller 2002) is that as a wholly avian virus,
no partial resistance may have existed in the human population
(Cox and Subbarao 2000). H5N1, to which human vulnerability
is universal, may be taking the same route of adaptation via a
smoldering avian progenitor (Shortridge et al. 2000).
Apart from record human case fatality, H5N1 has exhib-
ited other unique features unprecedented for avian influenza
viruses: unpreadapted lethality to laboratory primates (Rim-
melzwaan et al. 2003), ferrets (Govorkova et al. 2005), and ro-
dents (Gubareva et al. 1998), with cases of strong neuroviru-
lence (Maines et al. 2005); a mammalian host range includ-
ing feline (Keawcharoen et al. 2004), canine (Butler 2006b),
mustelid (Klopfleisch et al. 2007), and viverrid (Roberton et al.
2006) species; the establishment of distinct lineages all lethal
for gallinaceous poultry, but not necessarily waterfowl (Sturm-
Ramirez et al. 2005), and each with repeated transmission to
humans (Smith et al. 2006); a seeming increase in thermal and,
therefore, potential environmental stability (Swayne and Beck
2004; WHO 2006c); and geographic spread to the majority of
countries in Eurasia (OIE 2007). The emergence of H5N1, how-
ever, is but one example of the myriad appreciable changes that
seem to be taking place in the ecology and epidemiology of
avian influenza (Webster et al. 2007).
Since the first recorded outbreak of H5 subtype highly
pathogenic avian influenza (HPAI) occurred in 1959, 24 primary
HPAI outbreaks have been reported. In the first 20 years of this
47-year interval, 1959 to 1978, four outbreaks were recorded,
involving a combined total of less than 100,000 birds. In the
next 20 years, 1979 to 1998, a tripling of the frequency of out-
breaks was recorded with 13 outbreaks. The last seven years saw
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
264
M. GREGER
outbreaks of seven different HPAI viruses involving, on average,
millions of birds each (Alexander 2007). With the exception
of the 1983-1984 outbreak in Pennsylvania, which involved 17
million birds at an estimated cost exceeding 400 million USD
(Lasley 1987), all of the first 12 of 24 outbreaks recorded since
1959 involved fewer than 500,000 birds, whereas 8 of the last 12
outbreaks involved more than 500,000 birds and 7 of the last 12
involved millions of birds. Even excluding H5N1, which has led
to the deaths of hundreds of millions of birds across 44 countries
and territories (OIE 2007), outbreaks in the last decade of HPAI
viruses other than H5N1—namely H5N2, H7N1, H7N3, H7N4,
and H7N7—have affected more than 50 million birds across five
continents (Alexander 2007).
Historically, chickens have been considered a rare host for
influenzavirus A. Prior to the 1990s, few influenza viruses were
isolated from chickens and even fewer chicken-adapted lineages
described (Liu et al. 2003). This pattern dramatically changed
by the mid-1990s, as illustrated by the emergence, spread, and
eventual establishment of a stable H9N2 lineage in domestic
chickens throughout Eurasia (Xu et al. 2007).
Prior virologic survey studies made only 7 isolations of 3
subtypes of influenza viruses from 1,708 chickens sampled in
a Hong Kong poultry processing plant between 1975 and 1979
(0.4%), whereas 564 influenza viruses of 11 different HA sub-
types were isolated in ducks (6.5%) (Shortridge 1982). Chick-
ens were considered such atypical hosts that although virologic
surveillance in pigs in China continued, at the time human H5N1
infection was first recognized in 1997, the estimated 1,000 live
poultry markets in Hong Kong were not under surveillance
(Shortridge 1999).
In Hong Kong survey studies in the late-1970s, H9N2 viruses
were isolated from waterfowl but not from chickens (Shortridge
1982). The earliest H9N2 isolate in a terrestrial bird was discov-
ered in Hong Kong quail in 1992 (Guo et al. 2000). By 1995,
H9N2 was causing outbreaks of severe respiratory disease in
chickens in China (Guo et al. 2000) and Korea (Mo et al. 1997b)
with mortality rates of 10–40%. Currently, H9N2 is widely en-
zootic in chickens in Eurasia (Xu et al. 2007) and has infected
children in Hong Kong SAR in 1999 (Lin et al. 2000) and 2003
(Butt et al. 2005).
The unprecedented isolation of other influenzavirus A sub-
types in gallinaceous poultry has also been demonstrated re-
cently. Since 1995, H6 and H3 viruses have been isolated from
terrestrial poultry in China (Liu et al. 2003), Korea (Choi et al.
2005b), Taiwan (Lee et al. 2006), Italy (Campitelli et al. 2002),
South Africa (Abolnik et al. 2007), and, in the United States,
California (Webby et al. 2002), North Carolina, and Minnesota
(Choi et al. 2004). Particular concern surrounds the reported in-
crease in the prevalence in Hong Kong SAR of H6N1 viruses that
possess seven of eight genes highly homologous to the H5N1
virus that triggered the 1997 human outbreak in Hong Kong
(Chin et al. 2002). Surveillance data from 2000 to 2005 suggest
multiple co-circulating H6 viruses have become established in
gallinaceous poultry in southern China (Cheung et al. 2007).
Without more extensive pre-1990s surveillance data, it is
not possible to accurately assess how much of this escalation
is due to changes in detection or reporting, rather than actual
changes in incidence. Given the prior data that are available
on low pathogenicity avian influenza (LPAI) virus isolation
(Shortridge 1982), as well as the often conspicuous nature of
HPAI outbreaks (Capua and Alexander 2004), a true ecological
shift seems to have occurred in avian influenza viruses in the
1990s.
1b. Shifting Epidemiology. Before 1997, known direct
zoonotic infections with avian influenza viruses were consid-
ered rare and of little consequence (Subbarao and Katz 2000).
The four isolated case reports that did exist all described de-
tections of H7N7 viruses: the first from the blood of a patient
with hepatitis in 1959 (Campbell et al. 1970), followed by three
instances of conjunctivitis—a laboratory ocular splash in 1977
(Taylor and Turner 1977), an animal handler sneezed on by
an infected seal in 1981 (Webster et al. 1981), and a woman in
England reporting a piece of straw entered her eye while she had
been cleaning out her duck house in 1996 (Kurtz et al. 1996).
There is no convincing evidence that human infections with
H5 influenza viruses have ever occurred before H5N1 (Webster
et al. 2007). The scarcity and largely benign nature of the hu-
man cases reported before 1997 are supported by studies show-
ing that humans directly inoculated with avian influenza viruses
from waterfowl, the influenzavirus A reservoir, show at most
mild transitory infections in a minority of subjects (Beare and
Webster 1991).
In contrast, since 1997, more than 300 reported cases of hu-
man H5N1 infection have been confirmed across a dozen coun-
tries, resulting in nearly 200 deaths, the slim majority of which
were in children or young adults under the age of 20 (WHO
2007). Paralleling the ecological changes in a multiplicity of
avian influenza subtypes since the 1990s, H5N1 is just one of
five poultry influenzavirus A subtypes that have caused zoonotic
infections since 1997: H9N2 infections in children in China in
1999 (Peiris et al. 1999) and 2003 (Butt et al. 2005); H7N2 in-
fections in residents of New York in 2002 and Virginia in 2003
(CDC 2005b); H7N7 in poultry handlers and their family mem-
bers in the Netherlands in 2003 (Fouchier et al. 2004); H7N3
infections in poultry workers in Canada in 2004 (Tweed et al.
2004) and likely Italy in 2003 (Puzelli et al. 2005); H10N7 in-
fections in infants in Egypt in 2004 (PAHO 2004); and, in the
United Kingdom, H7N3 in 2006 (Nguyen-Van-Tamet al. 2006)
and H7N2 in 2007 (HPA 2007).
In 1993, children in Europe were reportedly infected with
an H3N2 avian/human reassortment virus likely transferred
from pigs (Claas et al. 1994). Combined with the unprecedented
human/avian/porcine triple reassortment in North Carolina pigs
in 1998 (Zhou et al. 1999) after a stable “classical” swine flu
lineage had presided throughout the bulk of the century (Brown
2000b), this suggests that the ecology and epidemiology of
swine influenza viruses have been changing since the 1990s
as well.
Critical Reviews in Microbiology Downloaded from informahealthcare.com by National Medical Library on 12/22/10
For personal use only.
THE HUMAN/ANIMAL INTERFACE
265
Aside from the H5N1 panzootic, the avian influenza out-
break in the Netherlands in 2003 is likely the largest in history
(Alexander 2007). Overshadowed by the burgeoning SARS epi-
demic at the time (Anderson et al. 2004), 30 million chickens
were culled or killed (Du Ry van Beest Holle et al. 2005) and
clinical infections were confirmed in 86 poultry handlers, in-
cluding an attending veterinarian who later died in acute respi-
ratory distress, and 3 of their family members (Fouchier et al.
2004). Human-to-human transmission was suspected due to the
infections in household contacts, and, indeed, a later government
investigation found serological evidence of infection in 33 of 56
(58.9%) household members of infected poultry workers with
no known exposure to infected birds (Du Ry van Beest Holle
et al. 2005) with plausible second-generation human transmis-
sion (van Boven et al. 2007). This suggested an ease of intrafa-
milial avian influenza transmission on par with seasonal human
influenza (Jennings and Miles 1978). Based on the seropreva-
lence of H7 antibodies, it was estimated that at least 1,000 and
perhaps as many as 2,000 people were infected in the outbreak
(Bosman et al. 2005).
In contrast with the four case reports of single human in-
fections recorded prior to 1997, in the last decade hundreds of
zoonotic avian influenza cases have been reported across four
continents. In the Netherlands outbreak, apart from the one fatal
case, most of the confirmed human H7N7 infections were rela-
tively benign conjunctivitis or a mild flu-like illness (Fouchier
et al. 2004), but human-to-human virus transmission seemed
extraordinary. H5N1 has the inverse relationship—high case fa-
tality but extremely limited human transmissibility. The 2003
Netherlands outbreak suggests avian influenza viruses may now
directly, without recombination or adaptation, result in efficient
human-to-human transmission; the human H5N1 cases starting
in 1997 suggest that avian influenza viruses can directly be viru-
lent zoonotic pathogens. Should the escalation of avian influenza
outbreaks and human infections continue, both characteristics—
extreme virulence and human transmissibility—may combine
to produce a human pandemic of grave proportions. Describ-
ing the acceleration in size, scope, and frequency of highly
pathogenic avian influenza virus outbreaks in recent years, Ilaria
Capua told Science:“We’ve gone from a few snowflakes to an
avalanche” (Enserink 2005). To slow this apparent rapid, re-
cent emergence of zoonotic poultry influenza viruses, it must
first be understood what triggered this “avalanche” in the first
place.
Integrated rice-duck farming (Shortridge 1992), pig-hen-
fish aquaculture (Scholtissek and Naylor 1988), and live poul-
try markets (Woo et al. 2006) have all been implicated in the
emergence of avian influenza viruses with pandemic potential.
The subsequent dissemination of viruses like H5N1 has been
blamed on factors such as the songbird (Melville and Short-
ridge 2004), exotic bird (Borm et al. 2005) and cockfighting
trades (Gilbert et al. 2006), migratory wild birds (Weber and
Stilianakis 2007), and the trade in commercial poultry and poul-
try products (Peiris et al. 2007). None of these putative risk fac-
tors for emergence and spread are new, however. Domesticated
ducks have been used as an adjunct to rice farming in Asia
for centuries (Shortridge 2003a), fighting cocks (Scott 1983)
and poultry have been traded for millennia, and wild birds have
been migrating for millions of years (Proctor and Lynch 1993).
The recent intensification of the global poultry sector may bet-
ter account for the “complete revolution” (Alexander 2007) in
the ecology and epidemiology of avian influenza over the last
decade.
1c. Global Poultry Sector Intensification. Peiris et al.
(2007) implicates the expansion of intensive poultry husbandry
for facilitating the increasing scale and frequency of HPAI out-
breaks around the world starting in the 1