ArticlePDF Available
Mapping of poverty and likely
zoonoses hotspots
Zoonoses Project 4
Report to Department for International Development, UK
The objective of this report is to present data and expert knowledge on poverty and
zoonoses hotspots to inform prioritisation of study areas on the transmission of
disease in emerging livestock systems in the developing world, where prevention of
zoonotic disease might bring greatest benefit to poor people.
Delia Grace, International Livestock Research Institute, Kenya
Florence Mutua, International Livestock Research Institute, Kenya
Pamela Ochungo, International Livestock Research Institute, Kenya
Russ Kruska, Consultant
Kate Jones, Institute of Zoology, UK
Liam Brierley, Institute of Zoology, UK
Lucy Lapar, International Livestock Research Institute, Vietnam
Mohamed Said, International Livestock Research Institute, Kenya
Mario Herrero, International Livestock Research Institute, Kenya
Pham Duc Phuc, Hanoi School of Public Health, Vietnam
Nguyen Bich Thao, Hanoi School of Public Health, Vietnam
Isaiah Akuku, ILRI intern
Fred Ogutu ILRI intern
Report sumbitted 18th June 2012
2.1Poverty............................................................................................................................... .............21
3.4Outbreakzoonoses.......................................................................................................................... 55
Mapping and measuring the burden of zoonoses, the density and number of poor livestock keepers
and emerging markets for livestock products can help identify the ‘hotspots’ where zoonoses not only
impose significant burdens but where zoonoses management is likely to be pro-poor (targeted at poor
livestock keepers and poor consumers of livestock products) and have most impact on helping small
farmers reach emerging markets.
All zoonoses are not equal and a first step of the study was to categorise zoonoses according to
epidemiology and impact. We considered three groups of zoonoses:
Endemic zoonoses are present in many places and affect many people and animals.
Outbreak or epidemic zoonoses are sporadic in temporal and spatial distribution.
Emerging zoonoses newly appear in a population or have existed previously but are rapidly
increasing in incidence or geographical range. Many occur as outbreaks.
The first chapter reviews the substantial literature on prioritising disease and identifies prioritisation
criteria relevant to this study, namely: burden of human disease; impacts on livestock production and
productivity; amenability to agricultural intervention; and, concern because of emergence or severity.
This allowed us to identify 24 zoonoses of high importance to poor people, 13 of which we
investigated in depth. Our priorities were broadly similar to comparable exercises.
The next chapter reviews current evidence on poverty and livestock, on livestock systems and their
dynamics, and on zoonoses and how they are currently mapped. We update the map of poor livestock
keepers of Thornton et al. (2002) and present an additional map based on sub-national data. Maps of
livestock systems that are changing most rapidly in response to emerging markets are taken from
Herrero et al 2009 and Notenbaert et al 2009), and vulnerability to climate change from Ericksen et al
2011). The strengths and weaknesses of different maps are noted and quantitative examples provided
on the massive under-reporting of zoonoses and animal diseases in poor countries.
The next chapter presents evidence from a systematic review of over 1,000 studies on the prevalence
of the 13 priority zoonoses in people and animals. It focuses on the endemic zoonoses that impose
greatest burden and a ‘top 20’ list is given of geographical hotspots. Data on zoonoses are also
extracted from the WHO Global Burden of Disease and the ‘top 20’ countries identified. We include a
case study that compares our systematic review with an ‘in-country review’ focusing on grey literature
and literature in a language other than English. Finally, we discuss some of the challenges of the
study and caution in interpreting the results. Maps are presented.
The next chapter updates the map of emerging disease events of Jones et al. (2008). For the first
time, we map emerging zoonoses as distinct from other emerging disease events. A ‘top 20’ of
geographical hotspots is given. Maps are presented. The last chapter provides maps of regional agro-
ecosystems and summarises numbers of livestock, people and poor livestock keepers by system as
well as the zoonoses context. It also draws some global conclusions from the study.
Annexes provide references for the papers in the systematic review of endemic zoonoses, the in-
country review, and the systematic review of emerging zoonotic events. They provide information on
the long list of zoonoses and the selection of the 13 most important to poor people in terms of burden
and economic impacts.
There is a strong association between poverty, livestock keeping, and zoonoses
Strength of evidence: strong
Zoonotic disease has many aspects and existing disease reporting systems do not adequately capture
the impact of zoonoses or identify investment opportunities. There is much unpublished information in
grey literature of developing countries.
Strength of evidence: strong
Across a range of zoonoses burden, poverty burden, and reliance on livestock, the hotspots for
poverty, emerging livestock systems and zoonoses are (in decreasing order of importance both by
region and country; countries in red appear in multiple listings):
South Asia: India > Bangladesh > Pakistan
Is higher than: East and Central Africa: Ethiopia > Nigeria > Congo DR > Tanzania > Sudan
Is higher than: South East Asia: China > Indonesia > Myanmar > Vietnam
Is higher than: West Africa: Burkina Faso > Mali > Ghana
Strength of evidence: moderate
We updated maps of poor livestock keepers (table 0.1). Around 70% of the rural poor and 10% of the
urban poor are dependent on livestock. The last decade has seen declines in density of poor livestock
keepers in South America and South East Asia and lesser declines in parts of West Africa and South
Asia. High density of poor livestock keepers is focal: around 6 hotspots and 14 countries bear the
brunt. Four countries (India > Nigeria > Ethiopia > Bangladesh) have 44% of poor livestock keepers
(table 0.1 and chapter 4.1).
Strength of evidence: moderate
Areas with both high livestock populations and strong rising demand for livestock products offer
highest opportunities for livestock to be a pathway out of poverty. Demand is largely driven by
urbanisation, demographic growth and increasing wealth. Monogastric (poultry and pig) production
responds more to increased demand because of their high reproduction rates and ease of
intensification. Hence total number of monogastrics and magnitude of change in monogastric
population are proxies for identifying emerging livestock systems (table 0.1 and chapter 4.3).
Countries with both high numbers and large change include: India > Myanmar > Pakistan >
Bangladesh = China
Strength of evidence: moderate
The study distinguishes between three categories of zoonoses:
Endemic zoonoses, present in many places and affecting many people and animals are
responsible for the great majority of human cases of illness (we estimate 99.9%) and deaths (we
estimate 96%) as well as the greatest reduction in livestock production. Examples are: brucellosis,
leptospirosis, and salmonellosis. Endemic zoonoses are of most concern where the objective is
lowering the burden of human disease and increasing the productivity and profitability of livestock
for poor people.
Outbreak or epidemic zoonoses are zoonoses that typically occur as outbreaks. Examples are
anthrax, rabies, Rift Valley fever, and leishmaniasis. They are much more sporadic in temporal
and spatial distribution than endemic zoonoses but may be more feared because of their
unpredictability and in some cases, severity. They are often present in neglected populations with
poor health services and infrastructure. Outbreak zoonoses are of concern when there is an
objective of reducing vulnerability of neglected populations.
Emerging zoonoses newly appear in a population or have existed previously but are rapidly
increasing in incidence or geographical range. Many occur as outbreaks. They are relatively rare,
around 300 events in the last 70 years. Most are of minimal impact, but historically, emerging
diseases have been responsible for massive impacts (e.g. HIV AIDS). Emerging zoonoses are of
concern when the object is foresight, and understanding disease emergence in order to try and
avert pandemics of major impact.
Strength of evidence: strong
The study assessed 56 zoonoses, together responsible for around 2.5 billion cases of human illness
and 2.7 million human deaths a year. We identified the 13 zoonoses most important to poor livestock
keepers because of their impacts on human health, livestock sector, amenability to agriculture-based
control, and other criteria (chapter 2). These were, in descending order: zoonotic gastrointestinal
disease; leptospirosis; cysticercosis; zoonotic tuberculosis; rabies; leishmaniasis; brucellosis;
echinococcosis; toxoplasmosis; Q fever; zoonotic trypanosomosis, hepatitis E; and anthrax.
Strength of evidence: moderate
The study searched for papers on zoonotic disease emergence events since 2004. Out of 43 new or
newly identified events, most are viral and zoonotic from wild animal hosts. Although the mappable
zoonotic new events (n = 30) are globally spread across every continent, there may be clusters in
northeast US, South America, Europe and South East Asia. These trends may reflect surveillance
differences and the possible trend to more events in developing countries may reflect increased
attention over this period. Combined with existing data on zoonotic EID events from 1940-2004 (n =
202), the clearest potential hotspots are USA and Western Europe, (this may also reflect historical
surveillance differences). Countries with most events are USA, UK, Australia, and France (table 0.1).
Strength of evidence: weak-moderate
Massive under-reporting constrains our ability to understand and prevent disease. In sub Saharan
Africa, 99.97% of livestock losses do not appear in official reports. At least 50% of these losses are
probably due to notifiable diseases as farmers and experts rank many notifiable diseases as major
causes of mortality (Newcastle disease, African swine fever, classical swine fever, trypanosomosis,
East Coast fever, peste de petits ruminants most notably).
Strength of evidence: moderate to strong
The study accessed information around 1000 surveys on prevalence of endemic zoonoses, covering
over 16 million subjects. A qualitative and semi-quantitative analysis suggests a strongly spatial
distribution, with a few countries bearing most of the human and animal disease burden (chapter 3).
The study also assessed the burden of zoonoses in the Global Burden of Disease (GBD) extracting
data on 7 important zoonoses. This also shows a highly skewed distribution of human disease burden:
19 countries are responsible for 75% of the total burden. Hotspots are: Nigeria, Ethiopia, Tanzania,
Togo, and India.
Strength of evidence: moderate
Countries appearing multiple times at the top of multiple metrics are (in descending order of
importance): India, China, Bangladesh, Ethiopia, Nigeria, Pakistan, Congo DR, Indonesia,
Myanmar, and Tanzania1.
Table 0.1: The top 20 countries at the interface of poverty, emerging livestock systems and zoonoses
according to different metrics (in descending order of importance)
cs (TLU)
Rapid change
India India China Myanmar India Nigeria USA
Nigeria Ethiopia Brazil Burkina Faso Nigeria Ethiopia UK
Ethiopia Nigeria Indonesia India Congo DR Tanzania Australia
Bangladesh China India Pakistan China Togo France
Congo DR Congo DR Viet Nam Ghana Ethiopia India Brazil
Pakistan Bangladesh Iran Afghanistan Bangladesh Mali Canada
Kenya Pakistan Philippines Bangladesh Pakistan Vietnam Germany
Sudan Indonesia Thailand Liberia Afghanistan Sudan Japan
China Angola Nigeria Central African
Republic Angola Bangladesh China
Tanzania Afghanistan Ukraine Chad Brazil Burkina Sweden
Indonesia Tanzania Pakistan Cambodia Indonesia Cameroon Italy
Madagascar Brazil Myanmar Benin Niger Chad Malaysia
Niger Philippines Bangladesh Laos Tanzania Rwanda Switzerland
Uganda Uganda Peru Thailand Kenya Ghana Congo DR
Turkey Mali Colombia Zimbabwe Côte d'Ivoire Mozambique Sudan
Philippines Sudan Ecuador Ethiopia Uganda South Africa Argentina
Afghanistan Mozambique Morocco Guinea Sudan Congo DR India
Egypt Malawi South Africa Guinea-Bissau Burkina Egypt Israel
Mozambique South Africa Bolivia China Mali Gambia Peru
Burkina Viet Nam Egypt Mali Iraq Ivory Coast Trinidad &
Pakistan Uganda
Zimbabwe Vietnam
^ Protein-energy malnutrition is the term used in the WHO GBD. WHO defines it as a nutritional
deficiency resulting from either inadequate energy (caloric) or protein intake and manifesting in either
marasmus or kwashiorkor.
*More than 20 countries because of tied ranks
Data sources for table 0.1. Poor livestock keepers, this study; Protein energy malnutrition: extracted
from WHO GBD, 2009; Monogastrics, number of poultry and pigs in developing countries converted to
tropical livestock units (TLU), FAOSTAT, 2012; Rapid change monogastrics: % increase in pigs and
poultry in developing countries from 2000 to 2030, this study (based on IMPACT model); Zoonoses
burden GBD: Burden of zoonoses extracted from WHO GBD, 2004 using the assumptions set out in
chapter 3; Endemic zoonoses prevalence, this study; Emerging zoonoses events, this study,
Density of poor livestock keepers (update of Thornton et al., 2002 by Kruska, this study)
This map shows the density of poor livestock keepers (number of poor livestock keepers per km
square). Countries with most poor livestock keepers are ranked in table 0.1 and estimates of the
absolute number of poor livestock and density of livestock keepers per country are given in chapter 4.
Hotspots for poverty & livestock keeping
Absolute numbers: India > Nigeria > Bangladesh > Congo > Pakistan > Kenya > Sudan
High density: South Asia, East Africa highlands, great lakes, Nigeria, west Africa, littoral
South-East Asia
Key points
Around 1 billion poor people (<$2 a day) depend on livestock
Around two thirds of the rural poor and one third of the urban poor depend on livestock
Livestock provide one fifth to one half of household income for the poor
In poor countries, livestock provide from 6 to 36% of protein intake
More details on poor livestock keepers are provided in chapter 2.
Change in poultry numbers from 2000 to 2030 – a proxy for emerging markets (Herrero et al.,
This map shows the percentage change in poultry numbers from 2000 to 2030 based on projections
by Herrero et al (2009) using the International Model for Policy Analysis of Agricultural Commodities
and Trade (IMPACT) model for the ‘business as usual’ scenario used in the International Assessment
of Agriculture Science and technology for Development (IAASTD). Increase in poultry production is
one possible proxy for the demand-driven increase in livestock production called the ‘livestock
revolution’. According to that study:
Hotspots for livestock sector growth in developing countries are:
Poultry in South and East Asia > bovines in South and East Asia > poultry in sub Saharan
Africa = pigs in sub Saharan Africa
Key points
Livestock production is increasing rapidly in response to growth in population, growth in
income, urbanisation and changing diets: the so-called livestock revolution
Herrero et al (2009), projects that, over the next 40 years, absolute growth in consumption will
be greatest in South Asia and South East Asia and relative growth greatest in sub-Saharan
On the supply side, growth will be greatest in the poultry sector followed by bovine then small
ruminants then pigs
Emerging livestock systems offer opportunities for smallholders if they can access the inputs
needed to reach emerging markets
More details on poor livestock keepers are provided in chapter 2.
Map of endemic zoonoses surveys – the burden of zoonoses (Grace et al., this study)
This map shows the surveys on endemic zoonoses reviewed in this report. Over 1,000 surveys were
accessed covering over 14 million animals, humans and livestock products.
Hotspots for high prevalence of endemic disease confirmed by multiple surveys
Nigeria, Ethiopia, Tanzania, Togo, India, Mali, Vietnam, Sudan, Bangladesh
Key points
In poor countries as a whole:
12% of animals have recent or current infections with brucellosis, reducing production by 8%
10% of livestock in Africa are infected with trypanosomosis, reducing their production by 15%
7% of livestock are currently infection with tuberculosis (TB), reducing their production by 6%
and from 3-10% of human TB cases may be caused by zoonotic TB
17% of smallholder pigs show signs of current infection with cysticercosis, reducing their value
and creating the enormous burden of human cysticercosis
27% of livestock show signs of current or past infection with bacterial food-borne disease, a
major source of food contamination and illness in people
26% of livestock show signs of current or past infection with leptospirosis reducing production
and acting as a reservoir for infection
25% of livestock show signs of current or past infection with Q fever, and are a major source
of infection of farmers and consumers
Map of emerging zoonotic disease events from 2004 to 2011 (Jones et al., this study)
This map shows locations of zoonotic emerging disease events between 2004 and 2011.
Geographical hotspots
Combined with existing data on zoonotic EID events from 1940-2004 (n = 202), the clearest potential
hotspots are USA, South America, South East Asia and Western Europe, which may reflect historical
surveillance differences.
This chapter identifies three categories of zoonoses important for different reasons: endemic
zoonoses, outbreak zoonoses, emerging zoonoses, and, old zoonoses. We discuss previous work to
prioritise zoonoses and some of the challenges in identifying zoonoses of importance to the poor. We
identified 56 zoonoses that appeared in multiple listings and selected criteria to prioritise them.
Together, the 56 zoonoses are responsible for an estimated 2.7 human million deaths and around 2.5
billion cases of human illness a year. For the top 13 zoonoses, the figures were 2.2 million human
deaths and 2.4 billion cases of illness. Our prioritisation is broadly compatible with other exercises.
1.1 Introduction
Zoonoses are diseases transmissible between animals (domestic and wildlife) and humans. Around
60% of all human diseases and around 75% of emerging infectious diseases are zoonotic (Taylor et
al., 2001; Woolhouse et al., 2005). In aggregate, they have high impacts on human health, livelihoods,
animals and ecosystems. In the first global syntheses of the impact (partial) of zoonotic diseases,
Grace et al. (2011a) estimated that, in least developed countries, 20% of human sickness and death
was due to zoonoses or diseases recently jumped species from animals to people.
1.2 The rationale for prioritising zoonoses
What cannot be measured cannot be managed and the first recommendation of a high-level WHO-
convened group was to assess the societal burden of disease attributable to zoonoses (Molyneux et
al., 2011). Assessing, including mapping, of zoonoses is key to guide decision-makers and
Zoonoses can threaten human health in different ways:
Endemic zoonoses are typically present to a greater or lesser degree in certain populations.
Examples are cysticercosis, brucellosis, bovine tuberculosis, leptospirosis and food-borne
zoonoses2. They are common in poor populations and are responsible around a billion
illnesses and millions of deaths every year (table 2.1). However, endemic zoonoses have
been neglected by the international donor, standard setting, and research communities. Maps
exist for human health burden of individual zoonoses, usually at country level but there are no
available maps of endemic zoonoses as a group, and few maps for the impact of zoonoses on
Outbreak or epidemic zoonoses typically occur as outbreaks. Examples are anthrax, rabies,
Rift Valley fever, and leishmaniasis. Endemic zoonoses may occur as outbreaks in naïve
populations or when triggered (e.g. sleeping sickness, leptospirosis). Outbreaks may be
triggered by climate changes, flooding, and epidemiological phenomena such as waning
immunity or other drivers. They typically have high temporal and spatial variability. Their
overall impact in terms of morbidity, mortality and production loss is much less than endemic
zoonoses but because they can ‘shock’ systems they are often of high priority to farmers and
decision makers. They can also cause important economic losses, which are often related to
reaction to the disease rather than the disease itself (Butler and Grace, forthcoming).
Emerging zoonoses newly appear in a population or have existed previously but are rapidly
increasing in incidence or geographical range. They are relatively rare, around 300 events in
2 Some of these also occur as outbreaks but are differentiated from the outbreak zoonoses in that community surveys will
generally show that the disease is present in communities, although it may only get attention when there is an outbreak
involving multiple cases.
the last 70 years (Jones et al., 2008). Most are of minimal impact. Many are endemic in other
places and burdens are not necessarily linked to site of emergence, so mapping the point of
emergence may not correlate to impact on poor people at that point. Donors and decision-
makers are often concerned about emerging diseases, whose impacts on poor farmers are
orders of magnitude less than the impacts of endemic zoonoses. However, the potential
impact (e.g. a new HIV AIDS) is at least of similar magnitude to endemic zoonoses. Good
maps exist but may not be useful for informing research aimed at identifying poor at risk from
‘Old zoonoses’ were originally zoonotic but are now spread mainly or entirely by human-to-
human transmission (some with zoonotic reservoirs) (Grace and McDermott, 2011). These
include HIV-AIDs, influenza, malaria, measles and dengue. These diseases can emerge
anywhere in the world and their burden is not linked to site of emergence. Their current order
of magnitude is about similar to that of the endemic zoonoses (almost all due to HIV-AIDs).
The ability of the livestock sector to predict, prevent and control these diseases is small and
maps not likely to be useful for directing research activities so these will not be further
1.3 Review of zoonoses prioritisation exercises
In order to map zoonoses and poverty, information is needed on which zoonoses pose risk to the
poor. One of the earliest attempts to prioritise zoonoses was conducted by ILRI (Perry et al., 2002)
with support from DfID. Recent years have seen several other prioritisation exercises for zoonoses
and animal health. These have been reviewed by the ENHanCE group (
(Enhance, undated). Most use experts, criteria setting, and weighting to come up with lists. However,
when evidence is both highly scarce and highly scattered (as is the case for zoonoses) then expert
opinion is less useful; this is illustrated by major discrepancies between different systems of prioritising
(Perry et al., 2009).
Prioritisation exercises use various criteria including: livestock pathogens with a high actual human
disease burden; rare zoonotic pathogens with severe disease manifestations in people; arthropod-
borne and wildlife associated pathogens which may pose a severe risk in future (Havelaar et al., 2010;
health consequences in animals, economic consequences in animals (Dufour et al., 2006); public
health (severity and occurrence in humans), animal health (severity of disease coupled with economic
consequences and occurrence in animals), and food (occurrence in food) (Cardoen et al., 2009).
Most of these prioritisation exercises were done in rich countries and additional challenges in
identifying the zoonoses that matter most to the poor include:
Capturing multiple impacts: Many endemic zoonoses and some emerging zoonoses have
impacts on livestock causing death and reduced productivity as well as costs for control.
Lack of evidence on the adverse impacts caused by disease: Many zoonoses are not
notifiable so are not recorded in official statistics. Even for notifiable diseases, many national
reports are highly unreliable. As described in the previous chapter, under-reporting is a
serious problem both in animal and human populations.
Variability: zoonoses are often focal and some vary from year to year (predictably or not). For
example, Rift Valley fever may be absent for decades before causing severe problems.
Human trypanosomosis currently affects only thousands of people, but historically there have
been major epidemics affecting millions of people.
1.4 Selection of zoonoses for prioritisation
In order to select ‘important’ zoonoses for further study, we used information from five listings of
priority zoonoses or priority diseases that included zoonoses and were relevant to developing
1) The World Health Organisation Global Burden of Disease
2) The World Animal Health Organisation list of notifiable zoonoses
3) Zoonoses important to poor people identified by expert consultation (Perry et al., 2002)
4) The Rosetta listing of infectious causes of death
5) A systematic review of zoonoses commissioned by DFID, which identified 373 zoonoses as
important (Grace et al., 2011).
Zoonoses that appeared in more than one list were considered (n=56). We ranked these 56 zoonoses
according to criteria considered important by the authors of this study. We selected the following
criteria (Table 2.1):
Human mortality (>1,000 deaths per year
Human morbidity (>1 million people affected
High impact on livestock sector
Amenability to agriculture-based control
Emergence or severity of disease in people
The major difference between our criteria and criteria used in previous studies was the inclusion of
‘amenability to agricultural intervention’. The rationale was that the ability to do something about a
problem was an important criterion for prioritisation. The complete table and weighting used is given in
annex 1. By these criteria, 13 zoonoses were defined as most important (Table 1.3). These 13 were
selected for in-depth systematic literature review and mapping.
Together, the 56 zoonoses are responsible for an estimated 2.7 human million deaths and around 2.5
billion cases of human illness a year. For the top 13 zoonoses, the figures were 2.2 million human
deaths and 2.4 billion cases of illness. Nine of the top-ranked zoonoses were considered to have high
impact on livestock and all of the top-ranked zoonoses are amenable to agriculture-based
Table 2.1 The most important zoonoses in terms of human health impact, livestock impact, amenability to agricultural interventions, severity of disease and
emergence (data from WHO and authoritative literature: when several authoritative estimates the mid point is given)
Disease Wildlife
interface Deaths
1 million
Gastrointestinal (zoonotic) Important 1,000,000 800,000,000 2 1 1 1 0 5
Leptospirosis Very important 123,000 1,700,000 2 1 1 1 0 5
Cysticercosis Sometimes 50,000 50,000,000 2 1 1 1 0 5
Tuberculosis (zoonotic) Sometimes 100,000 554,500 2 0 1 1 0 4
Rabies Important 70,000 70,000 2 0 0 1 Severe 4
Leishmaniasis Important 47,000 2,000,000 2 1 0 1 0 4
Brucellosis Sometimes 25,000 500,000 2 0 1 1 0 4
Echinococcosis Not important 18,000 300,000 2 0 1 1 0 4
Toxoplasmosis Important 10,000 2,000,000 1 1 1 1 0 4
Q fever Important 3,000 3,500,000 2 1 0 1 0 4
Trypanosomosis (zoonotic) Important 2,500 15,000 2 0 1 1 0 4
Anthrax Sometimes 1,250 11,000 2 0 1 1 0 4
Hepatitis E * Sometimes 300,000 14,000,000 2 1 0 1 0 4
Chagas Important 10,000 8,000,000 2 1 0 0 0 3
Chickungunya Very important 12,500 500,000 2 0 0 0 Emerge 3
Clostridium difficile disease Possible 3,000 300000 2 0 0 0 Emerge 3
Dengue fever Minor 20,000 50,000,000 2 1 0 0 0 3
Ebola Very important 500 800 2 0 0 0 Severe 3
Hanta disease Very important 1,750 175,000 2 0 0 0 Emerge 3
Avian influenza Important 77 145 0 0 1 1 Emerge 3
Bov. Spongiform Encephalopathy^ Sometimes 182 188 0 0 1 1 Severe 3
Psittacosis Important 2,250 22,000 2 0 0 1 0 3
Japanese encephalitis Possibly bats 11,000 40,000 2 0 0 1 0 3
Buffalo pox Not important Negligible Common 0 1 1 1 0 3
Rift Valley fever Important 45 150 0 0 1 1 Emerge 3
Note: high human mortality gets a double weight of as the most important criterion for many stakeholders. Total score = (human death x 2) + (humans affected) + (high livestock
impacts) + (farm intervention possible) + (other concerns: severe or emerging disease). The maximum possible score is therefore 6 and the minimum 0.
* Importance of zoonotic transmission not fully known ^ Not a problem in poor countries
1.6 Comparing with other assessments
ENhanCE (undated) reviewed 12 methods of disease prioritisation. Two were global (FAO/OIE and
WHO), one focused on Rajasthan in India, while the rest focused on developed countries. A variety of
methods were used: risk assessment approach, multi-criteria decision tools, and qualitative methods.
Together the studies reviewed covered animal diseases, human diseases, and zoonoses. Of the 99
diseases appearing in the ranking, 33 were zoonoses.
Zoonoses appearing in multiple listings according to the ENhanCE review, in declining order of
number of appearances, were:
Leptospirosis = rabies
Campylobacteriosis = tuberculosis = West Nile virus = toxoplasmosis
Listeriosis = anthrax = echinococcosis = E. coli infection = BSE = botulism
Cryptosporidiosis = Japanese encephalitis = Q fever = Rift Valley fever = tetanus
Out of the top 18 ranked zoonoses across the 12 studies, 17 appeared in our top 25 listing. The
exception was West Nile, which has been most problematic in the Americas. Of the 33 zoonoses, 24
appeared in our list of top 25 zoonoses, suggesting reasonable similarity given the different criteria
and focus.
A notable characteristic of these prioritisations is the high ranking given to common, food-borne
diseases (salmonellosis, campylobacteriosis, toxoplasmosis, listeriosis, toxigenic E. coli, and
cryptosporidiosis). Decision makers and implementers using unstructured prioritisation often focus on
classical zoonoses or emerging diseases rather than food-borne zoonoses (Grace et al., 2010).
1.7 Mapping zoonoses: strengths and weaknesses
There are three types of existing zoonoses maps: emerging disease event maps, disease report maps
and research-derived prevalence maps. A summary along with strengths and weaknesses are
presented in table 1.2.
1. Emerging disease event maps. Jones et al. (2008) have identified emerging disease events as “the
first temporal emergence of a pathogen in a human population which was related to the increase in
distribution, increase in incidence or increase in virulence or other factor which led to that pathogen
being classed as an emerging disease”. They identified 335 events between 1940 and 2004: 60% of
which are zoonotic. An updated map showing only zoonotic emerging disease events is presented in
Chapter 4.
2. Disease report maps. There are several systems for reporting disease outbreaks: these are
summarised in table 1.2. The most authoritative is the World Animal Health Organisation (OIE) World
Animal Health Information Database (WAHID) maps. HealthMap ( aggregates all
the major disease reporting systems and information sources. Members of the OIE (currently 178
countries) have a legal obligation to report certain diseases (currently 115). Maps are generated from
the information provided.
3. Prevalence maps. These are based on studies assessing the prevalence of zoonoses in livestock,
livestock products and people. Global prevalence maps exist for some individual zoonoses but data is
often at country level. Some zoonoses have been mapped using geo-spatial data – notably
trypanosomosis. The World Health Organsiation (WHO) Global Burden of Disease (GBD) has also
been mapped at country level.
Table 1.2 Zoonoses disease and disease outbreak reporting systems
included Source of
data Strengths Weaknesses
World Animal
Health Information
Database (WAHID)
33 Reports from
Notifiable so all
172 OIE member
states obliged to
Little reporting of
endemic disease by
developing countries
Animal Diseases
System (TADinfo)
2 (EMPRES) Reports from
Supported by
FAO. Resolution to
village level
Not widely used.
Access to data limited
to users
All animal &
All reports verified
by qualified
and timeliness
dependent on quality of
Global Public
Health Intelligence
Network (GPHIN)
reports of
public health
sources and
then verified
Real time reports
in 7 languages Developing countries
Not universally
accessible and costs
with use
Global Early
Warning and
Response System
19 zoonoses
of OIE, FAO,
Official reports Developing countries
HealthMap Human,
animal and
plant diseases
data from
many sources
Real time, most
1.7 Challenges in reporting systems for zoonoses in developing countries
The challenges of mapping the multiple burdens of zoonoses include:
Reporting systems cover only few of the important zoonoses. There are over 600 zoonoses
and around 100 of these are of some importance (Grace et al., 2011). However, WHO GBD
and OIE only cover 11 and 33 zoonoses respectively.
The GBD does not distinguish between zoonotic and non-zoonotic causes of disease (for
several diseases tuberculosis, schistosomiasis, gastro-intestinal disease the proportion of
disease attributable to zoonoses is not known).
Zoonoses are often confused with other diseases (e.g. malaria and typhoid) and this
misdiagnosis leads to systematic under-reported in the human health system.
OIE reporting grossly underestimates the importance of endemic zoonoses – see next section.
Emerging disease databases give little information on actual burden on poor people or which
diseases are likely to be problematic.
Meta-disease reporting systems (summarised in HealthMap) are only as good as the data
they aggregate.
An important conclusion of our study is that massive under-reporting of zoonotic (and other diseases)
in developing countries is a major impediment to understanding prevalence and impacts of disease
and developing appropriate control. We illustrate this with the examples of brucellosis in poor
countries and Q fever in Africa and also compare official reports of notifiable diseases with probable
mortality of livestock in Africa.
a) The case of brucellosis – a well-known, widespread, notifiable zoonosis
Brucellosis is an important disease of cattle, sheep, goats and pigs. It is also important zoonosis and
is notifiable to the OIE. In cattle, it can be suspected on clinical signs, as it causes late abortion with
characteristic lesions on the placenta. It also causes carpal hygromas, a very specific indicator of
brucellosis. Diagnostic tests are widely available and relatively inexpensive.
Commonly used tests for brucellosis detect antibodies produced in response to infection. Antibodies
tested for persist for several months (IgG) or several years (IgM). Positive tests (to both antibodies)
indicate the animal is currently sick, is chronically infected or has been infected in the last year or so.
Hence positive tests are roughly equivalent to annual cases.
Our review captured information from 241 community surveys (that is, surveys from the general
livestock community and not targeting high risk animals) of bovine, sheep and goat populations,
representing 475,968 samples. The prevalence for different regions is shown in Table 1.2. From the
number of ruminants, the prevalence of seropositive cases, and the relation between sero-positivity
and disease we can predict the number of cases of brucellosis a year. The discrepancy between the
number reported and the number predicted is several orders of magnitude. For example, for every 1
million cases in East Africa less than one case is reported to OIE. The situation is similar for other
diseases reported to OIE. When there are 999,999 missed reports for every one report, surveillance is
not fulfilling its purpose.
Table 1.3 Predicting the number of annual cases of brucellosis based on sero-prevalence and
comparing to the numbers reported to the World Animal Health Organisation
prevalence % Number ruminants Predicted cases
a year Cases reported
East Africa 8.2 257,377,760 21,104,976 12
West Africa 15.5 197,716,517 30,646,060 37
South Africa 14.2 59,806,724 8,492,555 6305
North Africa 13.8 57,629,367 7,952,853 1073
South Asia 16.0 683,181,040 109,308,966 156
South East Asia 2.9 21,247,586 616,180 164
b) The case of Q fever – a less well-known, difficult to diagnose, notifiable zoonosis
Q fever is an infectious disease of animals and humans caused by a species of bacteria (Coxiella
burnetii). The main reservoirs are sheep, goats and cattle. It is highly contagious to humans and
typically causes influenza-like illness, although some infections are asymptomatic and in rare cases
fatal complications can ensue.
Q fever is a notifiable disease and appeared in the top 13 zoonoses in terms of impact on human
health, livestock sector and other criteria in our listing (table 2.1).
Most tests for Q fever detect antibodies. Antibodies may persist for several years. Most of the surveys
in our review were community based. For these, a positive result indicates current infection, chronic
infection or infection in the last few years.
In our review the average sero-prevalence from community surveys in Africa was 26% suggesting half
a billion animals infected each year.
We reviewed cases of Q fever reported to OIE between 2006 and 2010, retrieving 742 reports from 54
African countries. Only one report had numbers of animals affected, no report had population at risk.
The reports were:
Disease outbreak report 1 report, 8 animals affected, 1 death
Confirmed infection without clinical signs 10 reports
Disease present but without quantitative data 16 reports
Diseases suspected but not confirmed 10 reports
Disease absent 226 reports
No information available 479 reports
Given that surveys carried out in the field suggests millions of cases occur in livestock in Africa each
year, and that all surveys conducted in Africa found evidence of sero-positive animals indicating
infection is present. It is obvious that official reporting seriously under-estimates the occurrence of this
important notifiable zoonosis.
c) Comparing probable livestock mortality with notifiable disease reports
The World Bank and OIE produced a very useful atlas summarising animal disease reports between
2006 and 2009 (World Bank, 2011). This also allows us to assess under-reporting for developing
countries. We do this by estimating number of livestock in Africa from FAOSTAT, annual mortality from
systematic reviews, proportion of mortality likely to be due to notifiable diseases from expert opinion,
and we compare these with official reports to OIE.
Number of livestock in Africa
FAO estimate that globally there are 24 billion livestock in 2010 (FAOSTAT, 2012), corresponding to
2.4 billion livestock standard units (using the OIE definition given in the aforementioned Atlas) (World
Bank, 2011). Sub-Saharan Africa has two billion livestock corresponding to 253 million standard
livestock units.
Number lost each year
Numerous studies on African livestock indicate annual mortality is high. Otte and Chilona (2002)
reviewed production parameters of ruminants in traditional and non-traditional production systems
reported in published and grey literature between 1973 and 2000 (table 1.3). Depending on species
and age category, mortality ranged from 6-28% with three quarters of the species-age categories
having a mortality of 10% or more.
Table 1.4 Annual mortality (%) in traditional ruminant systems
Young Growing female Growing male Adult female
Cattle 22% 7% 9% 6%
Sheep 27 10 11 11
Goats 28 13 14 12
From Otte & Chilona, 2002
Otte and Chilona did not include poultry in the review but production parameters and characteristics of
family poultry production have been compiled and published for eleven African countries (IAEA 2002).
These give a range of mortalities from around 30% to 80% depending on the age category and
country. Rege and Gibson (2009) estimate mortality in backyard poultry in Africa at 70%. There is little
comprehensive information on mortality among smallholder pigs, but mortality is often high among
pre-weaned piglets in smallholder systems (around one fifth) and very high losses occur during
outbreaks of African swine fever and other epidemics (Wabacha et al., 2004).
Proportion of losses due to notifiable disease
Some of the annual livestock losses are due to non-infectious causes (mainly accidents, poisoning,
predation and malnutrition). Other losses will be due to non-notifiable diseases (such as
endoparasites) but farmers and experts agree that the 87 notifiable diseases are among the most
important causes of mortality for livestock in Africa (e.g. Newcastle disease, trypanosomosis, classical
swine fever, East coast fever, contagious bovine pleuropneumonia, and peste des petits ruminants) in
sub Saharan Africa. The authors of the report consider at least 50% of mortality is due to notifiable
Combining these assumptions indicate a major discrepancy between the probable losses from
notifiable disease (around 10 million) and the losses reported to the OIE (around 100,000) can only be
explained by under-reporting of several orders of magnitude. The following worked example makes
this clear:
Livestock in Africa = 253 million standard livestock units (FAOSTAT, 2011)
Livestock death, slaughter, or destruction reported to OIE = 82,319 units (World Bank, 2011)
Livestock annual estimated losses = 25,300,000 TLUs (literature: 10% as a conservative
Estimated losses due to notifiable diseases = 12, 800,000 TLUs (expert opinion: 50%)
Losses (notifiable) reported to OIE = 0.2% (less than one fifth of one percent)
Losses (notifiable) probably not reported to OIE = 99.8% of losses
3 Mortality is lowest in cattle which contribute the most to tropical livestock units so we chose a low
estimate of mortality
Information of reasonable quality is available on number of people in poverty by country and the
number of livestock by country and by farming system. Literature provides estimates on the number
and proportion of poor people keeping livestock. Emerging zoonotic disease events have been
mapped, but because of the nature of the data, maps may not be informative about the impacts on
poor people. The World Health Organisation information on the Global Burden of Disease provides
information by country on the human health impact of around 11 important zoonoses and also on
protein-energy malnutrition which is indirectly linked to livestock product availability and hence
zoonoses. The World Animal Health Organisation collates information on 33 zoonoses but data from
developing countries is prone to under-reporting.
2.1 Poverty
Poverty can be defined as a pronounced deprivation in wellbeing. No single indicator exists to
measure all dimensions of poverty simultaneously, however, internationally comparable metrics, such
as the US 1$ a day ($1.25), are useful for spatial and temporal comparisons. Estimates of poverty are
probably reasonably accurate. The proportion of people living in poverty (<$1.25 per day4) dropped by
half between 1990 and 2010, but 1.3 billion people still live on less than $1.25 a day and around 2.5
billion on less than two dollars a day (World Bank, 2012).
In the past 3 decades, dramatic drops in poverty are mainly due to development in China: in Africa
and South Asia numbers of people in poverty or stable or increasing. In terms of numbers, more than
75% of the people living in poverty live in 9 countries and 80% of poor people in 12 countries. In terms
of intensity of poverty, 17 countries have more than 50% of the population living on less than $1.25
per day. Whereas in 1990, nine tenths of the poor lived in poor countries, presently three quarters live
in middle-income countries (mainly India, China and Brazil).
2.2 Livestock
How many livestock are kept and where are they?
In 2012, the human population reached 7 billion and the production animal population around 24
billion (FAOSTAT, 2012). Global livestock systems have been recently re-mapped (Robinson et al.,
2011). Poultry and pigs increasingly dominate in terms of number of animals kept (although in terms of
tropical livestock units ruminants are more important): 85% of all domestic animals alive are now pigs
or poultry. As disease transmission is dependent on numbers and contact rates, and monogastrics are
kept in higher numbers and more intensive systems, monogastrics may become more important in
disease emergence.
Livestock density maps
Livestock density reflects the number of livestock and the level of intensification. Human population
density is a major determinant of livestock density. High density is also an important factor in the
transmission of disease through increasing the probability and number of contacts. However, density
may also be associated with better biosecurity and control systems which reduce risk. High livestock
density, especially of monogastrics, often reflects intensification and tends to be inversely correlated
with poverty. In our study, livestock density seems more correlated with zoonotic disease event
emergence than burden of zoonotic disease. Figures 1.1 to 1.3 show global cattle density from the
FAO gridded livestock maps (FAO, 2007).
4 2005 international prices
Figure 2.1 Global poultry density (Robinson et al., 2011)
Figure 2.2 Global pig density (Robinson et al., 2011)
Figure 2.3 Global cattle density (Robinson et al., 2011)
2.3 Poverty and livestock
How many poor people depend on livestock? Where are they?
Livestock keeping had been variously regarded as a symptom of being poor, an important pathway out
of poverty, and a transitional stage as burgeoning developing world populations shift from agriculture
to urban livelihoods (Perry et al., 2010). Recent estimates suggest nearly 1 billion people living on less
than two dollars a day are dependent to some extent on livestock (Staal et al., 2009). Over 600 million
are found in South Asia, mostly in India. Sub-Saharan Africa has over 300 million poor livestock
keepers, concentrated in East and West Africa, with fewer in southern and central Africa. A breakdown
by region is provided in chapter 4.
What proportion of the poor depends on livestock?
Earlier estimates were around 70% of the rural poor depended on livestock (LID, 1999). Others have
estimated that 40-50% of those living in poverty ($1.25 threshold) are at least partially dependent on
livestock (Thomas & Rangnekar, 2004; IFAD, 2004). A more recent 12-country study supports this,
finding that on average, around 68% of rural households in the bottom 40% as regards expenditure
kept some farm animal compared to 65-58% of those in the 40%; in urban areas 22-26% of the poor
kept livestock, and 8-12% of the well-off (Pica-Ciamarra et al., 2011).
To what extent do poor people depend on livestock?
Staal et al. (2009) analysed 92 case studies from the developing world and found that livestock
contributions made up on average 38% of household incomes (33% of the income in mixed crop-
livestock systems, and 55% of total income in pastoral systems). The 12-country study of Pica-
Ciamarra et al. found livestock contributed on average 12% to household income, with no statistical
differences between contribution in rich and poor countries. (This study over-represented emerging
countries, which may explain the lower contribution of livestock compared to the study of Staal et al.,
2009). There is strong evidence that poor people depend on livestock, but more research is needed
on the extent and nature of this dependence.
Livestock provide many benefits besides income. These include traction, manure, food, social status
as well as economic services such as insurance and guarantees. Several studies show that manure,
while seldom marketed is highly valued in smallholder systems (ranking higher than milk in West
Africa (Grace et al., 2009). A study in Kenya, found that non-marketed values comprised
approximately 20% of the animals total perceived value (Ouma et al., 2003).
How much do livestock products contribute to nutrition in developing countries?
Across a range of developing countries, livestock products contribute 6-36% of protein and 2-12% of
total calories (Nzuma & Randolph, 2008). In South Asia and East Africa dairy products account for
most livestock product consumption; in the rest of Africa, dairy, poultry, beef and shoats are balanced,
while in South East Asia poultry and pork predominate. Countries with low livestock consumption (e.g.
Bangladesh) may offset this with high fish consumption.
What livestock do the poor keep?
Poorer households are more likely to keep small ruminants and richer to keep large ruminants. Poultry
keeping tends to be evenly distributed across wealth groups. However, species ownership is system
and country specific.
Which livestock systems contribute most value in poor countries?
Around half the value of production in sub-Saharan Africa is derived from cattle (55%), followed by
poultry (25%) and small ruminants (20%). In South Asia these proportions were 61%, 21% and 18%,
respectively (FAOSTAT, 2012). In both regions the arid/semi-arid zone contributed most to value of
production. However, the trends are towards more value from poultry and pigs and more production
from intensified systems. A recent comprehensive rural poverty mapping is the CGIAR Geographic
Domain Analysis (2009). The most recent public domain maps on global poor livestock-keepers are
those produced by ILRI in 2002 (Thornton et al., 2002).
2.4 Mapping poverty and livestock systems
Understanding the spatial distribution of livestock keepers can guide the allocation of resources as a
first step in reaching the poor; identify areas of opportunity for livestock as a catalyst to growth; and,
target hotspots of potential livestock-associated disease and environmental degradation. However, our
knowledge of the location, characteristics and trends of change among poor livestock keeping
populations is very patchy, both spatially and temporally. Here we outline a rapid broad- brush global
assessment of spatial distribution of poor livestock keepers, and describe parallel activities in high-
resolution poverty mapping for countries in East Africa using sophisticated econometric techniques
pioneered at the World Bank.
In 2001 the UK Government’s Department for International Development (DFID) commissioned a
study to produce sets of maps locating the significant populations of poor livestock keepers in the
world, and to assess in very broad terms how these populations are likely to change over the next
three to five decades. These data are reported by Thornton et al (2002). The map presented in this
document is updated using more current poverty estimates from World Bank’s 2011 WDI with the
majority of the information being from 2005-2010 yet with a few countries with estimates still from the
1990s. Several countries within each region also have no estimates and have to be extrapolated by
The Thornton et al (2002) study made use of existing data and spatial data layers, together with
information from the literature and expert opinion. The central element of the analysis is a global
livestock classification based on that of Seré and Steinfeld (1996), who present a typology based on
mixed crop-livestock systems, livestock-only rangeland-based systems, and landless production
systems. We defined the classification primarily in terms of landuse/cover and climate-based length of
growing period (LGP), supplemented by existing global coverages of human population, irrigated
lands, and urban areas. Human population scenarios to 2050 were developed for Africa, Latin
America and Asia.
The study also provided a breakdown of poverty information by country and livestock production
system that was available for most of the countries only at the national level including: World Bank
rural and national rates and two internationally comparable poverty lines: less than 1$/day and 2$day.
But this information did not include any information on how many of these poor were livestock
keepers. So one additional layer was created by assigning differential poverty rates by broad livestock
systems (mixed, pastoral and other) within each country providing at least some further sub-national
distribution of the poor with livestock.
2.5 Maps of poor livestock keepers
The updated map of density of “poor livestock keepers” 2010 based on the methodology of Thornton
et al. (2002) is shown in Figure 2.4. There are many assumptions and extrapolations involved in map
development, however, despite caveats, various conclusions can be drawn from the analysis. In terms
of numbers of livestock keepers, the critical regions remain South Asia and sub-Saharan Africa. The
mixed farming systems (crop and livestock) contain large numbers of poor (over 1 billion), and
numbers of poor people dependent to some extent on livestock are considerable. Mixed rainfed
systems have more poor livestock keepers than mixed irrigated systems. Rangeland systems have
least absolute numbers of the poor but the poor in this system have highest dependency on livestock.
Almost half of the poor in rangeland systems are located in sub-Saharan Africa.
Figure 2.4 Density of poor livestock keepers in developing countries based on national data (updated
March 2012)
Some changes are evident since the map of 2002. There has been a marked decrease in the density
of poor livestock keepers in South America and SE Asia. There has been some improvement, but to a
lesser degree in francophone West Africa and South Asia
Change in number of poor livestock keepers
Because of different methods in developing the maps, the map of 2000 is not directly comparable with
updates. Some changes are evident since the map of 2002. There has been a marked decrease in the
density of poor livestock keepers in South America and South East Asia. There has been some
improvement, but to a lesser degree in parts of francophone West Africa and South Asia. However,
these improvments have been more than offset by increases in Africa, most of South Asia, the middle
East and Central Asia. Overall, the number of poor livestock keepers is estimated to have by 56
million in the eight years from 2000 to 2008 (FAO, 2011).
Figure 2.5 Density of poor livestock keepers as mapped in 2002
The International Food Policy Research Institute IFPRI (Wood et al. 2009) have released their sub-
national rural poverty rates that cover most of sub Saharan Africa, but not yet the rest of the
developing world. We used these to prepare Poor Livestock Keeper (PLK) maps for the < $1.25/day
and < $2/day poverty lines (Figure 2.6 and 2.7).
The two methods (national and sub-national) have broadly comparable results. In both maps, Ethiopia,
Nigeria, the Great Lakes region, parts of West Africa and Malawi have the highest density of poor
livestock keepers. However, the sub-national data reveals differences within countries. Because
endemic zoonoses and emerging zoonotic events are geo-located, maps derived from sub-national
data are more useful in exploring associations between zoonoses and poor livestock keepers (Figure
Figure 2.6 Density of poor livestock keepers (<$1.25 a day) in sub-Saharan based on sub-national
data (update May 2012)
Figure 2.7 Density of poor livestock keepers (<$2 a day) in sub-Saharan based on sub-national data
(update May 2012)
2.6 Maps of livestock system change
Livestock systems are changing rapidly in response to various drivers. Figure 2.8 shows estimates of
livestock systems in 2000 and 2030 (Kruska et al., 2003, Hererro et al., 2008).
Figure 2.8 Farming systems in 2000 and 2030
Herrero et al 2009 modelled growth in different livestock systems using the IMPACT model (Fig 2.9).
Figure 2.9 Changes in monogastric populations 2000-2030 (Herrero et al 2009)
Under the reference, or ‘business as usual’ scenario the hotspots in terms of rapid growth are in
descending order, poultry in South and East Asia > poultry in South America > bovines in South and
East Asia > poultry in sub Saharan Africa = pigs in sub Saharan Africa. Figure 2.9 percentage
changes in pig and poultry densities between 2000 and 2030 (Herrero et al 2009). These are used as
a proxy for emerging livestock systems. These use estimates from the ‘baseline scenario’, that is the
most probable development of the sectors.
Does changing livestock systems change the risk of zoonotic disease emergence?
The maps in Figure 2.9 show some of the geographical areas where change is most rapid; countries
which are in the top 20 for both high numbers of monogastrics and rapid change are found in South
Asia and South East Asia: Myanmar, India, Pakistan, Bangladesh, Thailand and China. Change in
livestock systems extends beyond intensification of poultry and pigs. Rapid change and growth is
often associated with erosion of the natural resource base. Globally, anthropogenic changes are
driving climate change with implication for livestock keeping and zoonoses. These three change
processes are discussed in this section: intensification, interface with wildlife and climate change.
Agricultural intensification and zoonoses
A systematic review of zoonoses at the livestock/wildlife interface recently commissioned by DFID
examined evidence for links between livestock intensification and disease emergence. Evidence is as
yet insufficient for definitive conclusions, and in some cases intensification is associated with less
disease, overall intensification is linked with disease emergence and spread. Selection, breeding and
management for increased productivity in livestock create host populations conducive to pathogen
evolution and persistence (through lack of genetic diversity, high numbers and contact opportunities,
stress-induced immunosuppression and other factors). This provides opportunity for "wild"
microorganisms to invade and amplify or for livestock pathogens to evolve to new and more
pathogenic forms. In addition, corollaries of intensification such as high livestock and pest densities,
extensive transportation networks, sale of live animals for food and pets, landscape modification, poor
waste management, and juxtaposition of agriculture or recreation with wildlife all contribute to
"emergence" and shifting virulence of diseases (Grace et al., 2011).
The literature review conducted for this survey (chapter 4) found that zoonotic food-borne pathogens
were markedly higher in poultry and pigs than in small ruminants and cattle. This suggests that as
monogastric systems expand, so may food-borne disease.
Agricultural intensification is likely to have different impacts on the key zoonoses depending on their
epidemiology. Probable impacts are discussed in chapter 3 and summarised here. Of the priority
zoonoses, 9 are likely to become more of a problem with intensification, 4 are likely to decrease and
for the remaining there is no clear link.
Table 2.1 Probable impact of intensification on priority zoonoses
Zoonosis Likely impacts of agricultual intensification
Gastrointestinal (zoonotic) Most gastro-intestinal zoonoses are food-borne and likely
to increase with intensification and associated lengthening
and branching of food supply chains. Many gastro-
intestinal zoonoses causle little visible signs in animals
reducing farmer incentives for control.
Leptospirosis Leptospirosis is associated with smaller farms, and
pasture-grazing especially with stagnant water.
Intensification may reduce prevalence.
Cysticercosis Associated with free-range, scavenging pigs.
Intensification will reduce prevalence.
Tuberculosis (zoonotic) Associated with larger farms and confined systems.
Intensification likely to increase.
Rabies No clear link. Most human transmission from dog bites or
Leishmaniasis No clear link. Transmitted by sandflies. Domestic dogs are
the most important reservoir.
Brucellosis Associated with larger farms and confined systems.
Intensification will increase. However, artificial
insemination, often associated with intensification, will
Echinococcosis Associated with feeding offal to dogs. More common in
extensive systems.
Toxoplasmosis Some evidence this is more common in extensive
systems. Associated with rodents.
Q fever No clear link.
Trypanosomosis (zoonotic) Intensification reduces risk by removing tsetse habitat and
wildlife hosts
Anthrax No clear link
Hepatitis E * Extent of transmission from pigs not clear.
Chagas Most associated with extensive systems
Chickungunya Associated with incursion into forest areas.
Clostridium difficile disease No clear relation. Present in farm animals but role in
transmission not clear
Dengue fever Most transmission anthroponotic: livestock systems no
clear role
Ebola Intensification around bats is a risk
Hanta disease Spread by rodents. Not farm associated
Avian influenza Associated high poultry density – link with intensification
not clear
Bov. Spongiform Encephalopathy^ Associated intensive systems
Psittacosis No clear link
Japanese encephalitis Associated with intensive rice systems
Buffalo pox No clear link
Rift Valley fever May increase with intensification and irrigation
Zoonoses with a wildlife interface
Fourteen of the ‘top 25’ zoonoses have important wildlife reservoirs across many regions, including 8
of the ‘top 13’ zoonoses, namely: gastro-intestinal zoonoses, leptospirosis, rabies, leishmaniasis,
toxoplasmosis, Q fever, trypanosomosis and anthrax. For some of the other zoonoses, wildlife may
play an important role in some epidemiological circumstances. For example, tuberculosis associated
with conservation areas in Tanzania and South Africa, brucellosis associated with buffaloes, and
hepatitis E and cysticercosis with wild pigs. Where a wildlife interface exists, zoonoses control is much
more complex (Grace et al., 2011). Other zoonoses in the ‘top 25’ but not ‘top 13’ with an important
wildlife interface are: Chagas, Chickungunya, Ebola, Hanta disease, avian influenza, psittacosis and
Rift Valley fever.
2.7 Climate change and zoonoses
There are several metrics for vulnerability to climate change (Cutter et al., 2009; Fussel et al., 2009).
Tropical African countries and Asian coastal countries are usually among the countries considered
most vulnerable to climate change. The CGIAR Research Program on Climate Change, Agriculture
and Food Security (CCAFS) commissioned ILRI/ CCAFS to conduct a rapid assessment across the
global tropics of the vulnerability of food security to climate change (Ericksen et al., 2011). The goal
was to identify ‘hotspot’ locations where climate change impacts are projected to become increasingly
severe by 2050 and food insecurity is currently a concern, using a range of indicators. The maps
mainly focused on change with implications for crop growth but some of these changes also have
implications for livestock production and disease.
Figure 2.10 and 2.11 predict areas where rainfall and flooding will increase. This is expected to
increase the risk associated with vector-borne zoonoses including tick-borne and mosquito-borne
diseases. It will also increase risk of bacterial pathogens associated with stagnant water and flooding
(e.g. leptospirosis, anthrax, cryptosporidiosis).
Figure 2.10 Areas where rainfall per day increases by 10% or more between 2000 and 2050 (Ericksen
etlal 2011)
Figure 2.11 Flood frequency (Ericksen et al 2011)
The next map aggregates different thresholds that can stress production, including flips in: growing
period, reliable crop-growing days, annual aveage temperature, annual average maximum
temperature, maximum temperature exceeds 30 centigrade, changes in variablitly in rainfall, and
increase in rainfall.
Figure 2.12 Number of climate thresholds that can stress production (Ericksen et al 2011)
In terms of exposure to multiple climate threats, southern Africa has the largest area exposed (across
Namibia, Angola, Zambia, Botswana, Mozambique and South Africa) with multiple threats, followed by
northeastern Brazil, Mexico, Guyana, Nicaragua, and small areas in Tanzania, Ethiopia, the DRC,
Uganda, India, and Pakistan, as well as the Middle East.
While studies are starting to emerge on the likely effect of climate change on human disease, and
changes in spatial dynamics of some animal diseases (e.g. blue tongue) are believed to be influenced
by climate change, there is still little strong evidence on impacts of climate change on zoonotic
disease in dynamic systems with multiple drivers. For example, as countries get warmer and wetter
biological mechanisms would suggest that many diseases increase. However, if countries
simultaneously get richer or invest more wisely in health care the net impact may be disease decrease
(Perry et al., 2011).
The effects of climate change on livestock and non-vector-borne disease have, with some exceptions,
received little attention. The climate-livestock-poverty nexus was reviewed by Thornton et al. (2008)
and this section is largely based on their findings. Climate change may affect livestock disease
through several pathways:
Pathogens: higher temperatures and greater humidity generally increase the rate of
development of parasites and pathogens that spend part of their life cycle outside the host.
Changes to wind can affect spread of pathogens. Flooding that follows extreme climate events
provides suitable conditions for many water-borne pathogens. Drought and desiccation are
inimical to most pathogens.
Vectors: vector-borne diseases are especially sensitive to climate change. Changes in rainfall
and temperature regimes may affect both the distribution and the abundance of disease
vectors, as can changes in the frequency of extreme events (outbreaks of Rift Valley fever
have been linked to ENSO, for example).
Hosts: some will be exposed to new pathogens and vectors as their range increases and
impacts can be severe. Climate stress (heat, inadequate food and water) can also lower
Of the 13 priority zoonoses, cysticercosis, tuberculosis, rabies, brucellosis and echinococcosis are
unlikely to show high climate sensitivity and food-borne zoonoses, leptospirosis and trypanosomosis
are likely to show high climate sensitivity. It is less clear how climate change will affect the
epidemiology of other priority zoonoses, although some evidence suggests there may be important
negative impacts.
Food-borne zoonoses: A recent extensive literature review concluded that campylobacteriosis
and salmonellosis were most likely to increase with air temperature; campylobacteriosis and
non-cholera vibrio infections with water temperature; cryptosporidiosis followed by
campylobacteriosis with increased frequency with precipitation; and cryptosporidiosis followed
by non-cholera vibrio in association with precipitation events. Listeria sp. was not associated
with temperature thresholds, extreme precipitation events, or temperature limits (ECDC,
Leptospirosis: Leptospirosis is considered one of the more climate sensitive diseases.
Flooding and heavy rainfall have been associated with numerous outbreaks of leptospirosis
around the world. With global climate change, extreme weather events such as cyclones and
floods are expected to occur with increasing frequency and greater intensity and may
potentially result in an upsurge in the disease incidence as well as the magnitude of
leptospirosis outbreaks (Lau et al., 2010).
Trypanosomosis: While climate will modify (generally decrease, but not everywhere) habitat
suitability for the tsetse fly, the demographic impacts on trypanosomosis risk through bush
clearance are likely to outweigh those brought about by climate change (Thornton et al.,
Q fever is transmitted in aerosols and climate change could affect survivability. Toxoplasmosis
has rodent hosts and rodent populations are sensitive to climate change. Climate change and
other environmental changes have the potential to expand the geographic range of the
vectors and leishmaniasis transmission in the future. Anthrax is often associated with a
combination of heavy rain and warm temperatures following a drought that encourages spores
to germinate. These extreme events will be more common with climate change.
We undertook a systematic literature review of the 13 zoonoses identified as important. Eight of these
are ‘endemic classical zoonoses’ that is, zoonoses that are typically present across a wide range of
communities at most times (although perhaps showing annual and inter-annual variability). In the case
of bacterial food-borne zoonoses, we identified five diseases, which ranked highest on a number of
recent assessments of impact (salmonellosis, listeriosis, toxoplasmosis, campylobacteriosis and
disease caused by diarrhoeagenic Escherichia coli). We also considered three epidemic or outbreak-
associated zoonoses (rabies, leishmaniasis and anthrax) and hepatitis E, an emerging disease, which
may have an important zoonotic transmission.
For each disease we present information on prevalence, epidemiology, geographical hotspots and
some key research questions and we estimated an ‘endemic disease burden’ score for all the
countries for which information existed.
3.1 Methodology for systematic literature review
We generated search terms that incorporated certain key words, identified and screened abstracts,
reviewed full papers and synthesised required information. Diseases were initially considered on the
basis of their appearance in the top 13 list of zoonoses generated as explained in the previous
chapter. These were: Taenia solium cysticercosis, leptospirosis, anthrax, brucellosis, echinococcosis,
hepatitis E, leishmaniasis, Q fever, rabies, toxoplasmosis, trypanosomosis, tuberculosis, and food
borne infections (caused by Salmonella spp., Listeria monocytogenes, diarrhoeagenic Escherichia
coli, and Campylobacter spp.).
PubMed ( and CABDIRECT ( were used in doing
the searches but also Google ( including the Google scholar. Unpublished
materials including student theses (mainly from the University of Nairobi) were accessed by visits.
Related articles appearing during the active searches in PubMed were also utilized in sourcing for
extra details, as well as from relevant references cited in the main papers reviewed. When available,
the cited original papers were retrieved, reviewed, and relevant information retrieved. This also applied
to major review papers providing a summary of the needed information and references for the original
papers available. We used the same data as summarized in the review paper if the original paper was
not available.
Search terms were formulated, by disease, and by country or region; different combinations, were
either relaxed or broadened to capture more articles or were restricted to refine or limit the number of
resulting articles. Different Boolean operators were used (including AND, OR, parenthesis) for specific
PubMed and CABDIRECT searches. Phrases guided by key words were used for the Google
searches. We also applied wildcard symbols, mainly * to broaden the results in some of the searches.
The first step involved screening the abstracts by title, abstracts not relevant for the project objectives
were left out. The searches were originally limited to the last 10 years, but we also considered old
studies if the search results were initially few. We considered studies conducted in Africa, South Asia
and South East Asia. Some studies from the Middle East were also included. Those abstracts that
were considered relevant (based on the title) were extracted into a word document and subsequently
reviewed by a second person. Full papers linked to the relevant abstracts were extracted and
reviewed. Prevalence information, if available, was extracted from abstract in cases where the full
paper could not be accessed. Sources providing no information on the number of samples / subjects
analysed were not considered- as the basis for the calculation of the prevalence estimates could not
be established. Also excluded were papers with missing geographical locations for the specific
An excel® database was developed to capture information extracted during the review process
(including the different search terms used). Variables extracted included: country where the study was
done / or where the results apply, geo-spatial location (the specific location or coordinates if given),
number of herds studied, number of samples analysed, the specific diagnostic test(s) done, subjects
(livestock species, food, humans), individual prevalence, herd prevalence, year data were collected
and a description of the study population. Where multiple surveys were reported in one study, each
survey was listed separately (e.g. if prevalence was estimated in cattle and sheep these were
considered as two different surveys each with an associated sample size, species and prevalence).
We distinguish between “community studies’ which are conducted in the community and can be
considered representative of it, and “high risk studies” which were conducted in high risk populations
(sick people in hospitals, malnourished children, cattle which failed ante-mortem inspection, samples
taken during an outbreak etc.). Maps were generated using data from community studies only.
We defined geographical hotspots as those which had a high prevalence confirmed in multiple
surveys. The number of studies needed to consider estimates reliable varied from pathogen to
pathogen depending on the number of studies available and is given in the results section of the
different diseases.
In order to estimate the ‘top twenty countries’ for endemic disease burden we collated the
geographical hotspots. We standardized scores for each disease so the country that had highest
prevalence had a score of ten, and so on. We then summed scores for each disease by country.
To develop the maps of endemic zoonoses we collected on-locational or descriptive data on zoonoses
using systematic literature review and details were documented in a MS Excel spread sheet. The
study area covered entire Africa and Asia (South central Asia and South East Asia). The constitution
of a spatially referenced database was performed by introducing locational or spatial data in the form
of coordinates into the spreadsheet. These coordinates were approximated from the ‘Geospatial
location’ section of the database and were sourced from existing GIS databases and occasionally from
websites such as Google maps and
The spreadsheet was then imported into the GIS software package ArcGIS v.10 (ESRI, Redlands).
This software package allows for the seamless linkage of MS Excel spreadsheets to the GIS by using
the coordinates columns, and these are imported as event data. The zoonoses locations were then
mapped and the column ‘prevalence’ from the descriptive data used to map the magnitude of the
prevalence as a percentage. For visualization purposes, mapping was done with a base map of
agricultural farming systems on the background.
3.2 Results for systematic literature review- endemic zoonoses
We conducted a systematic review of brucellosis, tuberculosis, leptospirosis, trypanosomosis,
cysticercosis, and Q fever and some bacterial food-borne diseases.
Brucellosis the deceptive disease – causes fever and occasionally chronic disease in people;
mainly abortion and infertility in cattle, shoats and pigs
Tuberculosis white plague – a major cause chronic illness in people, causes wasting and illness
mainly in cattle
Leptospirosis swamp fever – causes fever and occasionally jaundice in people and fever and
infertility in cattle and pigs; wildlife important reservoirs
Q fever the most contagious disease – causes fever and occasionally death in people, carried
by cattle, shoats, pets and wildlife, causes abortion in shoats
Cysticercosis pork worm – most common cause of adult-onset epilepsy in poor, pig-keeping
communities, leads to carcass condemnation in pigs
Trypanosomosis sleeping sickness – cause of acute and chronic illness in people, historically
caused severe epidemics, the most important disease of cattle in sub Saharan Africa; wildlife
important reservoirs
Bacterial food-borne disease – the forgotten zoonoses – major cause of gastrointestinal disease in
people; some but not all cause illness in animals. Several have wildlife interface
Echinococcosis cystic disease - a major cause of illness in people and loss in sheep and goats
from condemnation of carcasses
We obtained information from 1098 surveys covering around six million animals, ten million people
and six thousand food or environment samples. Endemic zoonoses impose an important burden in all
regions, although the distribution varies according to disease. Trypanosomosis is found only in sub-
Saharan Africa, cysticercosis is rare (though not absent) from cultures where pigs are not kept, and
brucellosis is associated with high populations of ruminants. Zoonotic food-borne diseases, the most
important zoonoses, are at much higher prevalence in poultry and pigs than ruminants. Table 3.1
summarises the prevalence for important zoonoses by region and for all developing countries. It gives
the overall prevalence (humans, livestock, wildlife, other animals) and the prevalence for humans and
livestock separately.
Table 3.1 Prevalence (%) of important zoonoses by region
North Africa,
Near East
East Africa Southern
SE Asia All
Brucellosis* 13% 8% 14% 16% 16% 2% 12
Tuberculosis^ 9 8 5 7 17 0.2 7
Leptospirosis* 30 24 17 28 27 24 24
Q fever* 19 11 4 13 19 1 19
Cysticercosis^ Few pigs 12 23 16 14 12 14
Trypanosomosis^ Not present 9 12 10 N/A N/A 10
Food-borne disease 25 27 21 30 18 25 25
Overall 15 10 16 15 25 22 16
Human 15 15 11 10 19 11 16
Livestock 15 10 16 16 17 18 15
*based mainly on seroprevalence, indicates current or recent cases (last 1-2 years)
^ based on parasitological tests, indicates current infections
Brucellosis – the deceptive disease
The most important species of Brucella are zoonotic: B. abortus, responsible for bovine brucellosis; B.
melitensis, the main etiologic agent of ovine and caprine brucellosis and an increasing cause of cattle
brucellosis; and B. suis, causing pig brucellosis.
259 studies were assessed covering 476,067 animals and 31,842 people and 537 food samples. 248
studies were from communities and 11 from high-risk groups (mainly people in hospitals).
Commonly used tests for brucellosis detect antibodies produced in response to infection. A
combination of tests may be used to improve accuracy or ability to detect. The antibodies tested for
generally persist for several months (IgG) or several years (IgM). Positive tests (to both antibodies)
indicate the animal is currently sick, is chronically infected or has been infected in the last year or so.
Hence positive tests are roughly equivalent to annual cases.
In community surveys, the prevalence was 13% in shoats, 13% in bovines, 7% in camels, and 5% in
other species (chickens, pigs, dogs). Among livestock-keepers/abattoir workers prevalence was 11%,
and among suspect hospital patients, 7%. A large study in India found that 2% of patients in the
general hospital population tested positive for brucellosis.
The main risks for people are occupational (contact with livestock) and consumption of dairy products.
In some areas, brucellosis may be maintained in reservoir wild animal hosts (African buffaloes and
North American bison) in other cases diseases spills-over to wildlife and if eliminated in cattle
brucellosis will die out in wildlife. Brucellosis is more problematic in intensive systems than extensive
and pasture-based systems.
Hot spots
Brucellosis is mainly a problem where ruminants are important (Africa and South Asia). Shoat-keeping
communities are most at risk from B. melitensis considered the most pathogenic form.
Countries with multiple surveys (>=4) and high prevalence (>15%) include in descending order: Togo,
Mali, Ivory Coast, Zambia, Niger, India, Sudan, Cameroon and Burundi (human and animal combined)
Sero-positive animals have higher rates of abortion, stillbirth, infertility, calf mortality and lameness.
This is associated with lower milk yields (around 25% milk loss in aborted cows). Usually, infected
females will abort only once, although they may remain infected their entire life. The losses are
estimated at 6-10% of the annual value produced per animal (Mangen et al., 2002).
Agricultural losses have been estimated at $427 million per year for sub-Saharan Africa and $600
million for Latin America (Mangen et al. 2002; Seleem et al., 2009)
Human brucellosis usually presents as an acute febrile illness, often mistaken for malaria or typhoid.
Chronic complications are not uncommon.
Key research questions
Improving diagnosis in people, given widespread under-diagnosis and confusion with malaria
Public-private partnerships for control – promising studies suggest that by combining human
health investment and livestock sector investment, brucellosis can be controlled in a cost-effective
Role of wildlife in maintaining infection – wildlife have an important role in some circumstances.
The extent of this is not known, nor are effective strategies for managing disease in wildlife
Reducing risky behaviours around husbandry and consumption- much of the risk from brucellosis
can be reduced by simple precautions applied to handling cattle and food.
Developing a vaccine for B. suis
Effective vaccination which can be distinguished from infection to aid in control
Figure 3.1 Brucellosis prevalence in community surveys
Tuberculosis – the white death
Worldwide and historically, most human tuberculosis (TB) is caused by Mycobacterium tuberculosis.
M. bovis is responsible for cattle tuberculosis. It affects a wide range of animals and is responsible for
zoonotic TB in humans. In west Africa, M. africanum causes up to half of human tuberculosis – it has
characteristics intermediate between M. tuberculosis and M. bovis the agent responsible for bovine
tuberculosis. Atypical mycobacteria are found in the soil and environment and can infect both people
and animals.
110 surveys were assessed covering 336,152 livestock and 5,829 humans. 89 studies were
community-based and 12 in high-risk populations.
The standard method for detection of bovine tuberculosis is the tuberculin test, which involves the
intradermal injection of bovine tuberculin protein derivatives (PPD) and the subsequent detection of
swelling at the site of injection. This may be performed using bovine tuberculin alone or as a
comparative test using avian and bovine tuberculin. More recently, a gamma interferon test has been
developed. Meat inspection is also used to detect tubercular lesions in cattle, but in developing
countries is not very accurate. A large study in Ethiopia found routine inspection detected 3.5%
carcasses with lesions whereas detailed meat inspection procedures identified 10.2% carcasses, a
more than three fold difference (Biffa et al).
Positive tests are roughly equivalent to prevalence (or animals currently sick with TB).
Overall, 7.4% of livestock were positive. Overall prevalence was as follows: Bovines: 8%, camels
11%, shoats 2%, pigs 15%, wildlife 5%.
There were an estimated 12 million cases of human TB (prevalence) in 2010 (WHO, 2011). Twenty-
two high burden countries account for approximately 80% of all new TB cases. An extensive literature
exists on the prevalence of human TB but there is little information on what proportion is zoonotic, and
our review concentrated on this. Table 3.2 summarises more recent studies from developing
countries: on average 10.5% of human TB cases were associated with M. bovis. Our study suggests
a higher overall prevalence than previous best estimates (3.1%)5 but a strongly bimodal distribution:
zoonotic TB is either very important or minor in a given context.
Table 3.2 Studies since 1999 on proportion of zoonotic TB
Country Study MTBC M
bovis % M.
bovis Reference
abbreviation Year
Cameroon 15 district hospitals in Ouest 455 1 0.20% Niobe-Eyangoh 2003
Djibouti Unknown 85 1 1.20% Koeck 2002
Egypt Fever hospitals in cities 67 1 1.50% Cooksey 2002
Ghana Korle-Bu teaching hospital 64 2 3.10% Addo 2007
Guinea-b Unknown 229 4 1.70% KŠllenius 1999
Madagascar Antananarivo, Ansirabe,
Fianarantsoa, Mahajanga 400 5 1.30% Rosolofo-
Razanamparany 1999
Nigeria 2 hospitals Ibadan 60 3 5.00% Cadmus 2006
Nigeria Lagos 91 4 4.40% Idigbe 1986
Nigeria 3 hospitals Jos 50 10 20.00% Mawak 2006
Tanzania Arusha 34 7 20.60% Cleaveland 2007
Tanzania Pastoralist North & South 38 7 18.40% Kazwala 2001
Tanzania Arusha 34 7 20.60% Mfinanga 2004
Uganda Kampala 344 1 0.30% Asiimwe 2008
Uganda Kampala 234 1 0.40% Niemann 2002
Uganda Karamoja 10 3 30.00% Oloya 2007
Uganda Mbarara 69 0 0.00% Byarugaba 2009
Bangladesh Clinical 350 0 0.00% Nakajima 2010
India TB meningitis 37 24 64.90% Shan 2006
India EPTB hospital adjusted for prev
EPTB in population 155 22 2.90% Jain 2011
India EPTB hospital adjusted for prev
EPTB in population 115 53 12.60% Prasad 2005
Pakistan Hospital, Lahore 42 5 11.90% Nawaz 2012
MTBC= Mycobacterium tuberculosis complex EPTB=Extra-pulmonary tuberculosis
Prev. = prevalence
Cattle can also be affected by M. tuberculosis and can in turn shed this in secretions and excretions.
Cattle positive for M. tuberculosis to be a problem in South Asia: in one of the studies we reviewed,
taking place in India, 7.1% of pharyngeal swabs from cattle were positive for M. tuberculosis. This
5 Historically,Mboviswasresponsiblefor530%ofTBcasesintheUS,UKandNetherlands(OlmsteadandRhode,2011;
suggests that the burden of ‘zoonotic’ TB may be under-estimated as ‘human TB’ may be acquired
from cattle.
M. bovis infects a range of African wildlife (high levels of infection have been found in the Kruger Park,
Zambia and Serengeti). This is a potential source of infection to livestock and people, as well as a
threat to wildlife. Where M. bovis is established in wildlife hosts (e.g. badgers in the UK or possums in
New Zealand) eradication is very difficult.
Commonly found risk factors are close contact of animals (intensive and peri-urban systems),
increasing herd size and presence of wildlife reservoirs. Important risk factors for zoonotic TB people
are close contact with animals and consumption of raw milk. Prevalence appears to be higher on
intensive farms.
Muller summarizes a range of early reviews from Europe and North America before control was
widespread. Infected cattle lost 10% of milk production and 4% of meat production and infected cows
had one fewer calf. Unfortunately, good economic data is missing from developing countries but
similar losses could be anticipated. TB lesions are also an important reason for carcass condemnation
but it seems likely that routine meat inspection misses most cases (Biffa et al., 2010).
Zoonotic TB has a similar course in people as non-zoonotic. Overall, one third of the world’s
population is currently infected with TB. Of those infected with TB that do not receive treatment, about
5-10% will develop TB disease some time in their lives. Zoonotic TB is more likely to present as extra-
pulmonary, and prevalence of extra-pulmonary TB is a crude proxy for zoonotic TB.
Agricultural losses worldwide have been estimated at $3 billion (Garnier et al., 2003).
Geographical hotspots
Zoonotic TB is mainly a problem where cattle are important (Africa and South Asia). Dairying
communities are most at risk.
Countries with multiple surveys (>=2) and high prevalence (>5%) include in descending order:
Bangladesh, Burkina Faso, Pakistan, Ghana, Kenya, Mali, Cameroon, Chad, India, and Ethiopia.
Human TB (zoonotic 3-10%) is mainly localized in high burden countries: India, China, Indonesia,
Nigeria, Bangladesh, and Pakistan.
Key research questions:
M. tuberculosis (human TB) appears to be common in livestock in some areas (especially India).
What is the significance for transmission?
A zoonotic reservoir has been suspected for M. africanum – so far little evidence but not fully
TB is one of the most important and common human diseases. There is much uncertainty on the
proportion of this attributable to M. bovis, and our review suggested that the proportion is higher
than in previous estimates and most of a problem in South Asia
Impact of M. bovis on cattle in Africa and South Asia. Much of the information on impact is derived
from earlier studies in Europe or North America
Wildlife-livestock interface in hotspots (Tanzania, Ethiopia, Zambia and South Africa)
Understanding relation between intensification and disease: Cattle TB appears to increase with
intensification and urban farming
Controlling cattle M. bovis and M. tuberculosis in cultures that do not permit culling of cattle.
Figure 3.2 Tuberculosis prevalence in community studies
Leptospirosis – swamp fever
Leptospirosis is an infectious disease caused by pathogenic organisms belonging to the genus
Leptospira. There are many serovars (>250) but typically only around 10-20 are found in a given
region. Serovars can be grouped into 25 serogroups.
Microscopic agglutination test is the gold standard and was used by most of our studies. Paired serum
samples are used to identify current or recent infection. Antibodies may persist for several years. Most
of the surveys in our review were community based. For these, a positive result indicates current
infection, chronic infection or infection in the last few years.
109 surveys were assessed covering 52,534 animals and 83,596 people. In community surveys the
prevalence was 34% in swine, 29% in bovines, 14% in small ruminants, 16% in wildlife and 24% in
Among patients presenting with fever of unknown origins around 20% (7-57%) had leptospirosis.
Among patients with suspected leptospirosis 60-90% had positive diagnoses
Infected animals often become carriers. Wildlife are affected and can be important reservoir hosts.
Risk factors for humans include presence of rodents, farm animals and floods. Risk factors for animals
include smaller farms and extensive (pasture-grazing systems).
Geographical hotspots
Leptospirosis is mainly a problem in tropical countries where stagnant water can be found and where
cattle, pigs or rodents are frequent. SE Asia has been regarded as a hot spot but fewer studies have
been carried out in Africa.
Countries with multiple surveys (>=4) and high prevalence (>20%): Ethiopia, Vietnam, Nigeria, Egypt,
and Malaysia.
In people, leptospirosis most often presents as a febrile illness. Around 5-10% of cases may develop
jaundice or other complications and among these case fatality may reach 20%. SE Asia is considered
a hot spot and in some areas is the second most common cause of fever after malaria.
In livestock, leptospirosis is associated with abortion, still-birth, infertility and milk reduction in cattle
and swine.
There is little good data on losses associated with leptospirosis in developing countries. In Australia,
total loss was estimated at 2.2% at herd level (Holroyd, 1980).
In Vietnam, infection with some serovars correlated with one less live pig per litter, equivalent to 8%
loss of production (Boqvist et al., 2002).
Key research questions
Prevalence and incidence in Africa: leptospirosis has been considered most problematic in SE
Asia, this study suggests it may be more important than suspected in Africa
Leptospirosis as a misdiagnosis in people: like brucellosis and Q fever, leptospirosis is often
under-diagnosed and better tests as well as awareness raising among the medical community
and public is needed
Context specific vaccination – vaccination is effective but needs to be adapted for the serovars
Understanding whether livestock or wildlife are main reservoir: in some studies livestock
appear to be the most important reservoir, in others rodents. This has implications for control
Risk reduction: human behaviour is important in decreasing risk
Impact of climate change on extreme wet weather events and hence leptospirosis:
leptospirosis is strongly associated with flooding and stagnant water
Figure 3.3 Leptospirosis prevalence in community studies
Trypanosomosis – sleeping sickness and ‘the malaria of cattle’
Tsetse-transmitted trypanosomosis is an infectious disease unique to Africa and caused by various
species of blood parasites. The disease affects both people (Rhodesian and Gambian sleeping
sickness) and animals (nagana).
The most common test in animals is direct microscopic examination of blood for parasites. Hence
positive tests correspond to current infections.
103 studies were assessed covering 109,443 animals and 99,808 people. These were mostly
parasitological studies so they represent current infections. In community studies, there was a
prevalence of 10% among domestic animals and 5% among wild animals. Among humans (either
suspect hospital patients or in focal areas for trypanosomosis) prevalence was 6%.
The WHO reports human trypanosomosis as highly spatially distributed: in the last 10 years, over 70%
of reported cases occurred in the Democratic Republic of Congo (DRC). Angola, Central African
Republic, Chad, Sudan and Uganda make up most of the remaining burden.
Trypanosoma brucei gambiense (T.b.g.) is found in west and central Africa; it currently accounts for
over 95% of reported cases of sleeping sickness and causes a chronic infection (Gambian sleeping
sickness). Most transmission is anthroponotic and can be controlled effectively through interventions
targeted at human reservoirs; however, animal reservoirs have a role in the epidemiology. Pigs are an
animal reservoir and recently have been associated with the persistence and epidemics of sleeping
sickness in Uganda, Equatorial Guinea and Cameroon. Other domestic animals and wildlife are also
T. brucei rhodesiense (T.b.r.) is found in eastern and southern Africa. Nowadays, this form (Rhodesian
sleeping sickness) represents less than 5% of reported cases and causes an acute infection.
Agricultural expansion, deforestation and the removal of wildlife reduce the natural habitats and
wildlife hosts of tsetse. Moreover, applications of insecticides to cotton and other crops may also
reduce tsetse numbers and it is generally agreed that agricultural expansion/intensification is likely to
reduce trypanosomosis challenge, at least in the short term (Bourne and Wint, 1994). However, in the
short term there may be an upsurge in disease as tsetse, lacking alternative wildlife hosts, feed more
on cattle.
Acute sleeping sickness is a serious disease in people.
Trypanosomosis has serious health impacts in livestock. However, the non-zoonotic T. congolense
and T. vivax are less pathogenic than zoonotic T. rhodesiense. Swallow (1999) summarises a number
of studies and estimates reduced productivity of around 10-20% across a range of parameters
(mortality, calving rate, milk, draft power).
Geographical hotspots
Countries with multiple surveys (>=4) and high prevalence (>8%): Sudan, Mozambique, Tanzania,
Ethiopia, Cameroon, Nigeria, and Burkina Faso
Key research areas:
Arrest the northerly advance through Uganda of the zoonotic parasite T. brucei rhodesiense,
which threatens to converge with T. brucei gambiense
Farmer-and community-based management of disease: while technical highly effective,
sustainability remains elusive
Trypanocide resistance: an emerging problem across Africa which may also threaten the
efficacy of human drugs
Pen-side tests to allow better and more timely treatments avoiding cattle losses and slowing
development of resistance
Impact of climate, agricultural intensification and demography on disease dynamics
Factors leading to massive human outbreaks as occurred historically at the start of the 20th
century and to a lesser extent in the 1960s
Figure 3.4 Trypanosomosis (trypanosomiasis) prevalence in community surveys
Cysticercosis – pig tapeworm
Cysticercosis is a systemic parasitic infestation caused by the pork tapeworm (Taenia solium).
The tests commonly used for cysticercosis in pigs include meat inspection, lingual inspection and
antigen ELISA tests. These indicate current infections. In humans, stool samples are also used to
identify current tapeworm infections and imaging to identify brain cysts.
125 studies were assessed covering 349,923 pigs and 10,385,132 people. In community studies, the
average prevalence in pigs was 17%. Among humans the prevalence in community studies was 11%
(this combines people infected with Taenia solium as well as the much rarer cases of human
cysticercosis). Among hospital patients and epileptics the prevalence was 12%.
Humans are at risk not from consumption of pork with cysts but from consumption of tapeworm eggs
shed by themselves or another human carrier. The disease persists in poor, pig-keeping communities
where pigs have access to human faeces. Intensification would be expected to reduce prevalence of
the disease.
Hot spots: Rwanda, Congo, Chad, Togo, Nigeria, and Ghana (Geerts et al., 2004)
In some countries, pigs with visible infections (by lingual palpation or mucous membrane inspection)
fetch a lower price. This was estimated as a 30% reduction in price in the Cameroon (Praet et al.,
2003). A study in Tanzania estimated the price of healthy pigs at $45 and infected at $21: a reduction
of 46% (CIRAD, 2012). Heavily infected pigs may be condemned during meat inspection; however, in
many countries either smallholder pigs are not always inspected at slaughter or inspection is
Cysticercosis is believed to be the most common cause of adult onset epilepsy in poor, pig-keeping
In Cameroon, the cost of treatment of one cysticercosis patient (wage loss not included) was
estimated at Euro 260 (Praet et al., 2003)
Key research questions
Eradication of cysticercosis from an ecosystem: with the new vaccine as well as effective
therapeutics, cysticercosis is eradicable but there have been no serious investments in Africa
or Asia
Pen-side tests for diagnosis of cysticercosis in pigs: a lateral flow test has been recently
developed but requires serum (whole blood would be more convenient)
More comprehensive and effective meat inspection: as for TB, it appears that because of
financial incentives and dysfunctional systems, current meat inspection is not effective in poor
Hotspots among marginalised pig-keeping groups;
Figure 3.5 Cysticercosis prevalence in community surveys
Q fever – the most contagious disease
Q fever is an infectious disease of animals and humans caused by a species of bacteria (Coxiella
Most tests detect antibodies. Antibodies may persist for several years. Most of the surveys in our
review were community based. For these, a positive result indicates current infection, chronic infection
or infection in the last few years.
We accessed 81 surveys covering 27,252 animals and 11,023 people. In community studies,
prevalence was as follows: bovines 28%, other animals (cats, dogs, horses and poultry) 26%, shoats
15%. Among febrile patients in hospitals, 0-40% (average 8%) had antibodies to Q fever.
Coxiella burnetii is most frequently found in ruminants (cattle, sheep, and goats) but can also be
detected in wildlife and companion animals. According to the literature (although not in our review)
sheep appear to be infected most frequently, followed by goats and less frequently, cattle. Human
cases are often associated with proximity to small ruminants (particularly at parturition or during
abortions) and dry, windy weather. At least in Europe, here is no conclusive evidence in support of a
link between an increased density of animals and/or farms and spillover from infected farms to
humans (EFSA, 2010).
Geographical hotspots
Because Q fever has been investigated in few countries it is difficult to identify hotspots. Countries
with high sheep populations would be expected to be at higher risk. Countries with multiple surveys
(>=4) and high prevalence (>15%) include in descending order: Nigeria, Zimbabwe, India, and Egypt.
Animals that carry this organism and shed it into the environment usually do not show any signs of
disease. Infected ewes and does may abort or give birth to weak offspring. There is little data available
on the economic impacts.
In people around 50% of infections may be asymptomatic; other patients have influenza-like
symptoms, a minority have atypical pneumonia or hepatitis. In around 5% of patients, chronic infection
Key research questions
Prevalence studies in more countries
Economic impacts of Q fever in livestock
Factors leading to outbreaks of Q fever in human populations
High risk groups: pastoralists appear to have very high levels of Q fever – more studies are
needed on prevalence and prevention in this group
Q fever as an emerging disease
Vaccination to manage Q fever in high risk populations
Figure 3.6 Q fever prevalence in community surveys
Bacterial food-borne disease – the forgotten zoonoses
In this category we include the bacterial zoonotic diseases, which are transmitted mainly through food.
We reviewed Salmonella, toxigenic Escherichia coli, Listeria, Campylobacter and Toxoplasma which
are among the most important causes of food-borne disease as well as hepatitis E, an emerging
zoonosis. Other zoonoses of somewhat lesser importance not reviewed are: Staphylococcus aureus,
Bacillus cereus, and Clostridium spp. In this section, we do not include the previously considered
classical endemic zoonoses that are often food-borne (brucellosis, Q fever, zoonotic tuberculosis) but
have other important transmission pathways. We did not consider non-zoonotic diseases associated
with animal source foods (typhoid, rotavirus disease, scarlet fever, giardiasis, shigellosis etc.). We call
these ‘forgotten’ zoonoses because health experts, decision makers and the public are often unaware
of the important role zoonoses play in food-borne infections.
A variety of tests were used. In most cases, a positive result indicates current, recent or chronic
We accessed 258 surveys covering 27,425 animals and 263,995 people and 4,208 food or
environmental samples. In community studies, prevalence was as follows: bovines 16%, shoats 20%,
pigs 30%, poultry 36%, other animals 17%, food 31%. Among people in the community prevalence
was 21% and among high-risk groups prevalence was 20%.
Geographical hotspots
Countries with multiple surveys (>=4) and high prevalence (>15%) include in descending order:
Tanzania, South Africa, Gambia, Vietnam, Nigeria, Senegal, India, and Egypt.
Some of these diseases can have impacts in animals (salmonellosis, listeriosis, toxoplasmosis).
However, in many cases strains pathogenic to people are not pathogenic to animals which means
farmers have less incentives for control.
In people, food-borne disease is an important cause of illness and economic loss. There is no good
information on the proportion of gastrointestinal disease burden associated with food-borne zoonoses
in developing countries. In developed countries, the proportion varies from 30-50% (Grace et al.,
2008). Food-borne pathogens also cause other health problems, less common but more serious (e.g.
kidney failure, septicaemia, abortion, encephalitis etc.); around 2-3% of people with acute food-borne
zoonoses may also go on to develop serious complications (Lindsay, 1997). The health burden of
these is considered to at least equal the burden due to gastrointestinal illness.
Key research questions
Impact of food-borne bacterial diseases in livestock
Managing food safety in the informal sector where most of the poor buy and sell but food
safety regulation is not working
Gender and food safety – much of food purchase, processing, and handling is done by women
but they are often not engaged in food-safety programs
Relation between food safety and food security
Attribution – how much food-borne illness is due to zoonotic disease or agricultural products
High risk groups –the young, old, pregnant and immunosuppressed are especially vulnerable
to food-borne disease and special targeting is needed to reach them
Figure 3.8 Food borne disease prevalence in community surveys
Figure 3.9 Toxoplasmosis prevalence in community surveys
Echinococcosis – cystic disease
Cystic echinococcosis (CE) in humans is caused by the larval stage of E. granulosus, E. ortleppi, E.
intermedius or E. canadensis. All these parasites have canines (usually domestic dogs), as definitive
hosts and a variety of ungulates, particularly farm animals, as intermediate hosts. Man is generally an
aberrant intermediate host in which the hydatid cyst develops, usually in the liver or lungs as a space-
occupying lesion, which can result in considerable morbidity
We did not review echinococcosis in depth; however, a comprehensive assessment has recently been
carried out by Budke et al. (2006).
Cystic echinococcosis (CE) is a condition of livestock and humans that arises from eating infective
eggs of the cestode Echinococcus granulosus. Dogs are the primary definitive hosts for this parasite,
with livestock acting as intermediate hosts and humans as aberrant intermediate hosts.
Geographical hotspots
More than 90% of human cases occur in the 8 endemic regions in North Africa-near East and China.
In descending order: China (Tibetan plateau), Turkey, India, Iraq, Iran, and Afghanistan
A preliminary estimate of the annual global burden of CE has suggested approximately 1 million
DALYs are lost due to this disease (Budke et al., 2006). This is likely to be a substantial underestimate
(Craig et al., 2007). In addition the losses to the global livestock industry is around $2 billion lost
annually and cost of illness is around the same.
We conducted a systematic review for toxoplasmosis but have not included the information here
because of some epidemiological complexities. For toxoplasmosis, high prevalence may be
associated with less risk, because the most vulnerable group (pregnant women) are exposed before
they are pregnant. For hepatitis E there is some uncertainty over the extent of zoonotic transmission.
Key research questions
Control of echinococcosis in remote and insecure regions
Relation between toxoplasmosis prevalence and risk: as toxoplasmosis is most serious if
encountered by a naïve pregnant women, it may be that cultures where exposure to
toxoplasmosis is very high (e.g. France) have less disease burden than cultures where
exposure is low
Prevalence of toxoplasmosis
Changing behaviours that increase exposure to toxoplasmosis and echinococcosis
Zoonotic component of hepatitis E
3.3 Top twenty countries for endemic zoonoses burden
Twenty-eight countries appeared in the ‘geographical hotspots’ listing. To be considered a
geographical hotspot for a disease, a country had multiple surveys (human and animal combined, but
only community surveys) with a high average prevalence (table 3.2). (The cut-off prevalence varied
with disease reflecting that for different diseases, different prevalences are considered high6). The
country with the highest prevalence is ranked as 1. The ranking for each disease is shown in Table 3.2
as well as the number of surveys and cut-off prevalence the ranking was based on.
Table 3.2 Geographical hotspots for zoonotic disease (country)
Brucellosis Tuberculosis Leptospiros
is Cysticer
cosis Q fever Trypanosomo
sis Food-borne
1 Togo Bangladesh Ethiopia Rwanda Nigeria Sudan Tanzania
2 Mali Burkina Vietnam Congo Zimbabwe Mozambique South Africa
3 Ivory Coast Pakistan Nigeria Chad India Tanzania Gambia
4 Zambia Ghana Egypt Togo Egypt Ethiopia Vietnam
5 Niger Kenya Malaysia Nigeria Cameroon Nigeria
6 India Mali Ghana Nigeria Senegal
7 Sudan Cameroon . Burkina Faso India
8 Cameroon Chad Egypt
9 Burundi India
4+ surveys,
prev. >15% 2+ surveys,
prev. >5% 4+ surveys,
prev. >20% Geerts
et al. 4+ surveys,
prev. >15% 4+ surveys,
prev. >8% 4+ surveys,
prev. >15%
To calculate the countries with the highest burden of endemic zoonoses we gave each country a
weighting according to its ranking for average prevalence for each disease (the country with the
highest prevalence for the disease got a weighting of ten). Weights were added across the diseases
(so a higher score represents a higher average prevalence summed across all the endemic zoonoses
considered). The countries with highest burden of endemic zoonoses are shown in Table 3.3. (This
ranking is probably biased towards countries with better university and research infrastructure as they
conduct and publish more studies: for example, there are many more studies from Nigeria than from
the Central African Republic).
Table 3.3 Countries with most zoonotic disease hotspots
Country Score Country Score Country Score
Nigeria 27 Cameroon 10 Pakistan 8
Ethiopia 17 Chad 10 Zimbabwe 8
Tanzania 17 Rwanda 10 Zambia 7
Togo 15 Ghana 9 Kenya 6
India 14 Mozambique 9 Niger 6
Mali 14 South Africa 9 Senegal 4
Vietnam 14 Congo 8 Malaysia 2
Sudan 13 Egypt 8 Burundi 1
Bangladesh 10 Gambia 8
Burkina Faso 10 Ivory Coast 8
6 To give an extreme example, for a rare disease like rabies one in 100,000 animals might be
considered a high prevalence, while for a common disease like brucellosis one in 5 animals might be
considered high.
3.4 Outbreak zoonoses
We retrieved papers on three of the outbreak zoonoses that appeared in the ‘top 13’ zoonoses listing.
These were: rabies, anthrax and leishmaniasis.
Twenty-one papers were accessed on rabies. These were not useful for assessing prevalence or
cases but were consistent with the geographical patterns of rabies.
Most rabies cases are concentrated in high-risk countries in Africa and Asia. (Bangladesh, India,
Myanmar, Pakistan, China Egypt, Sudan, Ethiopia, Tanzania, Ghana).
Thirty-five papers were accessed on anthrax. They were not useful for assessing prevalence or cases
but were generally consistent with the geographical patterns of anthrax.
There are approximately 10-100 thousand human incidences annually throughout the world with
significant numbers of cases in Chad, Ethiopia, Zambia, Zimbabwe and India
Paper retrieval was not useful in assessing prevalence. Most leishmaniasis is zoonotic, but
anthroponotic transmission is more important in outbreaks, 90% of human visceral leishmaniasis
cases occurring in South Asia, Sudan, Ethiopia, and Brazil and 90% of cutaneous leishmaniasis cases
occurring in Afghanistan, Algeria, Iran, Saudi Arabia, Syria, Brazil, Colombia, Peru, and Bolivia.
Other important outbreak pathogens that were not in the ‘top 13’ list but did appear in the ‘top 25’
Rift Valley fever virus
Hanta virus
Ebola virus
Chickungunya virus.
Avian influenza was also in this list, a zoonosis that is endemic in some regions (Indonesia, South
China, Egypt and possibly elsewhere), but in most countries occurs as outbreaks, which are controlled
(rich countries) or burn out (poor countries) (Bett et al., in press).
These five pathogens (including avian influenza) are all caused by viruses and are characterised by
high case fatality but low burden of disease. Together they cause around 15,000 deaths a year which
is trivial in comparison to the top 13 zoonoses (causing 2.5 million deaths a year). Currently humans
are mainly spill-over hosts and there is no sustained anthropogenic transmission (human-to-human).
However, if these pathogens were to mutate to allow easy human-to-human transmission while
maintaining their high case fatality the impacts would be enormous. Hence, these diseases are of
interest not so much because of their burden of disease but because of potential to become diseases
with higher burden. Smallpox, bubonic plague, HIV-AIDS, malaria and measles are examples of
former zoonoses that jumped species with civilisation-altering impacts (Wolfe et al., 2007).
Perry and Grace (2009) argue that many negative impacts of zoonoses and emerging diseases are
from inappropriate responses by authorities, farmers and general public rather than disease itself. This
was especially evident in the avian influenza pandemic, when outbreaks led to large changes in
purchasing behaviour, which probably had little impact on mitigating risk. Similarly, the reluctance to
support commodity-based trade is prejudicial to developing countries without any commensurate
benefit in reducing human health risk.
The map was extracted from HealthMap ( It shows all disease reports between
Jan 1st 2010 and May 2nd 2012 of the endemic zoonoses considered in this report. The sources used
were: ProMed, FAO, OIE, Eurosurveillance, Google.
Figure 3.10 Outbreaks of five important zoonotic diseases 2010-2012 as aggregated and
reported by HealthMap (
(Extracted from HealthMap,
In all 200 reports are cited, on the HealthMap site, distributed as follows: leptospirosis (124),
trypanosomosis (24), brucellosis (20), Q fever (12) and bovine tuberculosis (10). Between Jan 1st 2010
and Dec 31st 2010 there were only three reports for brucellosis and all were in people. As explained in
Chapter 2 this implies under-reporting of actual new cases by several orders of magnitude.
3.5 Global burden of disease
The original Global Burden of Disease Study (GBD) was commissioned by the World Bank in 1991 to
provide a comprehensive assessment of the burden of 107 diseases and injuries and ten selected risk
factors for the world. Burden of disease is calculated using the disability-adjusted life year (DALY).
This time-based measure combines years of life lost due to premature mortality and years of life lost
due to time lived in states of less than full health. The GBD represents the most authoritative source of
information on human illness.
There are some challenges in using the GBD to assess zoonoses.
Firstly, zoonoses (especially in poor countries) are widely unreported, and under-reporting is
relatively greater for zoonoses than for non-zoonotic diseases of comparable prevalence
(Schelling et al., 2007). As the GBD report is based on national information for levels of mortality
and cause of illness, this under-reporting is reflected in the GBD.
Secondly, several zoonoses with considerable burdens are not included in the GBD assessment.
For example rabies, echinococcosis, cysticercosis, leptospirosis and brucellosis
Thirdly, the GBD is organised around diseases and not pathogens or transmission pathways. For
example, diarrhoeal diseases, among the highest causes of morbidity and mortality in poor
countries, comprise one category. Although the majority of important diarrhoeal pathogens are
zoonotic (Schlundt et al., 2004) it is not currently possible to identify the zoonotic component of
diarrhoeal disease from GBD figures
In order to use the GBD to estimate disease we made some assumptions:
Tuberculosis: we took the conservative estimate of Cosivi (1998) who estimated worldwide the
proportion of TB caused by M. bovis at 3.1%. This literature review suggests the proportion is
higher. A higher proportion is also consistent with historical data.
Diarrhoeal diseases: we assumed 33% of diarrhoea disease is due to zoonotic pathogens. In
developed countries, several reviews (Schlundt et al., 2004, Flint et al., 2005) argue the
majority of gastrointestinal disease burden is due to zoonotic pathogens (>50%). However,
given the lack of evidence for developing countries we took the conservative estimate of 33%.
Trypanosomosis, Chagas disease, leishmaniasis: Japanese encephalitis: are all considered
as zoonotic.
Schistosomiasis: cases in regions where the zoonotic species Shistosoma japonicum
predominates are considered zoonotic.
Dengue is not included. Although dengue is a zoonosis and the sylvatic cycle (monkey-
mosquito) has important implications for disease eradication, most transmission is human-to
Tetanus is not included. Tetanus is a sapro-zoonoses and the load of toxins in the
environment is largely the result of contamination with ruminant faeces. However, most human
burden is from contact with the environment and not animals.
Respiratory disease is a major cause of human sickness and death and a certain proportion is
due to zoonotic diseases such as Q fever. We did not include these as no reliable estimates
could be found.
For zoonoses recorded in the GBD, 68% of the burden is made up of just 13 countries (Figure 3.1).
There is a very high correlation (99%) between protein energy malnutrition7 and burden of zoonoses
indicating the strong relation between poverty, dependence on livestock, and zoonotic disease.
7 Protein-energy malnutrition is a nutritional deficiency resulting from either inadequate energy (caloric) or protein intake and
manifesting in either marasmus or kwashiokor. Marasmus is characterised by wasting of body tissues, particularly muscles and
subcutaneous fat, and is usually a result of severe restrictions in energy intake. Kwashiokor affects mainly children, is
characterised by oedema (particularly ascites), and is usually the result of severe restrictions in protein intake. However, both
Types can be present simultaneously (marasmic kwashiokor) and mask malnutrition due to thepresence of oedema.
Figure 3.11 Health burden of zoonoses in million disability adjusted life years (DALY)
From Global Burden of Disease, World Health Organisation, 2008
3.6 Comparing systematic literature review to in country literature search
Introduction and summary
In this review we extracted papers mainly in English (with a minority in French) from medical and
agricultural databases available on line. Systematic literature reviews which only include some
languages and which depend on major databases risk missing important information. Hence, we
conducted a study in Vietnam to review literature for three of the key zoonoses (Data provided by
154 papers were identified of which 117 were in Vietnamese and only 27 in English.
We used a large range of Vietnamese scientific journals, library documents, as well as meetings with
key researchers on zoonoses, and open sources.
•Vietnamese journals on preventive medicine, practical medicine, public health, veterinary sciences
and techniques, agriculture and rural development
•Institution libraries: Vietnam Medical Information Centre (MOH), National Institute of Hygiene and
Epidemiology, National Institute of Malariology Parasitology and Entomology, National Institute of
Animal Husbandry.
•University libraries: Hanoi Medical University, Hanoi School of Public Health, Hanoi University of
•Key researchers/research groups from institutes and universities
•Conferences proceedings
•Web sites: Ministry of Health, Ministry of Agriculture and Rural Development
The process of creating the search strategy consisted of two steps: (i) identification of key concepts
characterizing the research questions and (ii) generation of a list of search terms that reflected the key
concept. The main concept identified was zoonotic diseases in Vietnam. For this concept a number of
subject terms and keyword terms were identified, which was then combined for the search:
The overall search term components considered to define “zoonotic diseases” AND “Vietnam” for the
search, were: (i) population surveyed (human or animal), (ii) prevalence and (iii) laboratory techniques.
Diseases were searched by their common names, as well as the names of the causative agents. The
keywords were used both English and Vietnamese, for examples:
English: “Cysticercosis” AND “Vietnam” or “Taeniasis” AND “Vietnam”
Vietnamese: “Sán dây ln” or “Bnh ln go”
English papers: we searched online databases of Science Direct, Pubmed and Web of Science with
keywords of disease name or names of the pathogens in the fields of title/keywords/abstract.
All the electronic copies and hard copies papers were scanned and reviewed from their title and
abstract to see if the papers are relevant to research on zoonotic diseases in Vietnam. After 2
screening rounds, many zoonoses were identified from research done. However, due to the time
constraint and to respond to the TOR, we decided to select 3 zoonotic diseases, including
cysticercosis (pig tapeworm), leptospirosis, and Salmonellosis for in-depth review. For each of the
paper related to the selected diseases, we collected key information on i) location of the study, ii) on
human or animal or both, iii) robustness of research design, iv) analysis method, v) prevalence. It
happened also when a paper reporting different values for different sample analysis (e.g. milk, serum
or both), these were treated separately to have different prevalences of the targeted samples.
We found 50 papers, project reports and student’s thesis study related to cysticercosis, including 47 in
Vietnamese and 3 in English; 64 papers, project reports and student’s thesis study related to
salmonellosis, including 40 in Vietnamese and 24 in English; 40 papers, project reports and student’s
thesis study related to leptospirosis, including 30 in Vietnamese and 10 in English.
Conducting an in-country review covering Vietnamese journals, libraries and conference proceedings
dramatically increased the number of papers and samples. It also revealed papers on less commonly
studied aspects (e.g. cysticercosis in animals other than pigs and leptospirosis in wildlife) which were
missed by the systematic, web-based, mainly English language review.
For four of the six comparisons, the prevalence estimated by systematic and in-country review were
similar, but for two there was a marked discrepancy (27% prevalence of cysticercosis in people versus
4% and 57% prevalence of leptospirosis in livestock and companion animals (domestic) versus 14% in
the systematic and in-country reviews respectively). The much smaller number of studies in the
systematic review makes it likely that these are less accurate.
However, while 60% of the papers in Vietnamese were judged to have a ‘moderate’ or ‘weak’
methodology, only 37% of the papers in English were so judged.
Table 3.4 Comparing data on cysticercosis from a systematic review and an in-country review
Systematic review In country review
People Pigs Other People Pigs Other
Studies (number) 6 1 0 23 19 10
Community studies (number) 5 1 0 23 17 10
Prevalence (%) 27.4 9.9 3.8 10.8 24.6
Samples for prevalence (no.) 1,434 323 0 27,298 5,015,261 493,803
Table 3.5 Comparing data on leptospirosis from a systematic review and an in-country review
Systematic review In country review
People Domestic Wildlife People Domestic Wildlife
Studies (number) 1 3 0 58 167 5
Community studies (number) 1 3 0 43 166 10
Prevalence (%) 12.8 57.0 12.1 13.6 7.7
Samples for prevalence (no.) 961 456 0 7,085 22,506 1,047
Table 3.6 Comparing data on salmonellosis from a systematic review and an in-country review
Systematic review In country review
People Animals Food People Animals Food
Studies (number) 1 11 6 18 167 58
Community studies (number) 0 11 6 7 166 58
Prevalence (%) n/a 13.1 39.2 16 51.8 31.9
Samples for prevalence (no.) n/a 6831 980 2099 106,910 22,269
Figure 3.12 Leptospirosis and cysticercosis in Vietnam identified from in-country review
3.7 Interpretation of the review of endemic and outbreak zoonoses
As well as the likely existence of large amounts of missed literature because literature review was
based on English/French publications indexed on online databases, other weaknesses and potential
biases include:
Many zoonoses have never been looked for in many places, and published literature reflects
research infrastructure as well as disease prevalence
Zoonoses which are more difficult to detect or test for (e.g. campylobacteriosis, listeriosis) are
under-represented because they are rarely investigated
We generally used data from the last 10 years. For several zoonoses, information exists for 30
or more years. However, given rapid changes in farming systems as well as changes in
diagnostic techniques we thought this might be less reliable.
Some surveys don’t distinguish to species level making it impossible to distinguish between
zoonotic and non-zoonotic pathogens
Surveys focus on presence rather than transmission: so Mycobacterium tuberculosis will
usually be classified as non-zoonotic although it is possible that the human victim acquired the
infection from livestock
Most surveys only report on one pathogen
Varying sensitivity and specificity of surveys because of different tests used making direct
comparison difficult
Often little information on sampling and some community surveys may have considerable
selection bias although authors claim sampling was representative
In general, studies with fewer samples and country estimates based on fewer studies appear
to over-estimate prevalence. This may be because researchers focus on areas they think may
have a problem even if this is not always reported in the survey
Some areas are over-represented (near universities) and many under-represented
Some countries are under-represented because of less capacity in assessing zoonotic
In the current analysis small samples have as much weight as large samples. We planned to
construct weighted prevalences by sample size and extrapolate to populations at risk but time
did not permit (confidence intervals could also be given)
We class slaughterhouse surveys as ‘community’ that is representative of the livestock
population. In some places, sick animals are less likely to be slaughtered but in others they
are more likely
Suggestions for overcoming these problems and improving our understanding of zoonoses of
importance to the poor:
Further analysis of the data collected to allow better extrapolation to agro-ecosystems and
investigate risk factors
Conduct large scale, probabilistic, stratified surveys to accurately determine prevalence of key
Conduct surveys in regions where pathogens are likely to be present but data is lacking (risk-
Collect economic and behavioural data to understand the impact and risk factors for key
Develop better, cheaper diagnostics that can detect multiple pathogens
Support bio-repositories for pathogens with meta-data allowing investigation of epidemiology
and risk factors
The study searched for papers on zoonotic disease emergence events since 2004. Out of 43 new or
newly identified events, most are viral and zoonotic from wild animal hosts. Although, the mappable
zoonotic new events (n = 30) are globally spread across every continent, there may be clusters in
northeast US, South America, Europe and South East Asia. These trends likely reflect surveillance
differences with perhaps a trend to higher representation in developing countries reflecting increased
attention over this period. Combined with existing data on zoonotic EID events from 1940-2004 (n =
202), the clearest potential hotspots are USA and Western Europe, which may reflect historical
surveillance differences. Countries with most events are USA, UK, Australia, and France. (Chapter
prepared by IOZ)
4.1 Introduction
Novel pathogens continue to emerge worldwide, the majority of which are zoonotic. ILRI engaged
Zoological Society of London and Ecohealth Alliance to produce updated high resolution maps of
zoonotic human emerging infectious disease (EID) events, by 1) extraction from the database used in
Jones et al. (2008), and 2) further collection of EID events to update the database until 2012.
4.2 Methodology for updating map of emerging zoonotic disease events
We review the zoonotic human EID events in Jones et al. (2008), and methods for data collection and
mapping of new events.
1) Jones et al. (2008) EID database. The data from Jones et al. (2008) covers 335 human EID events
occurring between 1940 and 2004, 216 of which were zoonotic. In the Jones et al. (2008) analysis, an
EID event was defined as the first emergence of a pathogen in a human population as a result of
either increasing incidence, virulence, geographic range, drug resistance, or any other cited factor in
case reports and literature. Single case reports were excluded, as were those with uncertain data
quality. Separate pathogen strains were not considered as separate events, with the exception of drug
resistant strains. For the maps presented in Jones et al. (2008), the location of each EID event was
geo-referenced to point localities where possible, based on locations given in data sources (e.g. a
village, or a hospital). These data were then converted to give each pathogen a location in a one
decimal degree global grid. However, 95 of these events could only be traced to larger areas such as
subnational units (e.g. states) or whole countries. To standardise for further analysis in the paper, one
grid cell in these areas was selected via a random draw to represent the location of a particular EID.
One of the random draws was selected for the maps presented in the paper.
For the purposes of this report, we mapped all those events that had a point locality, and where only
subnational unit location information was available, we assigned an EID event to the weighted centroid
of that area calculated using the 'Generate Centroid' function in HawthsTools 3.27 for ArcMap 9.3.
Where only country information was present, these data were excluded from the maps. A total of 172
events were mapped (Fig. 1a).
2) Further data collection of EID events. We made slightly different assumptions when collecting new
data than in the original Jones et al. (2008) protocols, based on updated criteria and suggested
changes to the definition of “emerging” since 2008. We only considered emerging pathogens to be
those completely novel to humans, or having novel virulence in humans or novel drug resistance.
Geographic range, incidence and any miscellaneous factors were excluded. Furthermore, single case
reports are not excluded, in contrast to the previous map. Separate strains or subspecies are once
again only considered in the context of virulence and drug resistance. We have also extended our
criteria to cover issues not addressed by Jones et al. (2008). We excluded any events from non-
natural infections (for example, accidental inoculation of laboratory workers), and we accept novel
pathogens not yet given certified scientific names. Additionally, we explicitly definite an event as
zoonotic if there is evidence the pathogen has an animal host or vector, or if it is cited in the literature
as a likely zoonosis. Much assumption of zoonotic status is based upon transmission/natural hosts of
closely related pathogens. Zoonoses were sub-classified by host type into one of four groups: 1)
wildlife hosts only, 2) non-wildlife (i.e. domestic) hosts only, 3) both host types, 4) unknown. Finally,
we accepted any diagnostic method (e.g. serology, microbial culture, DNA) as evidence of human
infection (or animal infection for evidence as a zoonosis).
Initial leads on new EID events were obtained by various search methods including; 1) searches in
peer-reviewed journal articles using Web of Science v5.0 (search terms = [novel OR new] AND human
AND pathogen AND [emerging OR outbreak]); 2) searches in the ProMED reporting archives (search
terms = [novel OR new] AND human); and 3) expert suggestion (EcoHealth Alliance and colleagues).
Initial searches were conducted in articles from 2004 onwards, although no date restriction was made
in accepting EID events or their supporting references. From initial leads, further information was
found via follow-up of references. Once a potential EID event was identified, we extensively traced
literature on the pathogen both backwards and forwards to find the first chronological case (not the
first chronological report). Data were not recorded if the pathogen could not be assigned a single,
feasible spatiotemporal origin (for example, many novel human viruses have recently been discovered
that were then found in populations worldwide and/or historical samples, giving no clear emergence
time or location). Following Jones et al. (2008), we recorded multiple fields of interest including details
of pathogen taxonomy, transmission mode, known hosts, pathogenicity, data quality and any
miscellaneous notes, in addition to geographic location. This information was generally unavailable for
pathogens described very recently (in the past 10 years or so) because they have not yet been
studied in detail, though no exclusions were made based on missing supplementary data.
4.3 Results: new zoonotic emerging disease events
We identified 43 new EID events, 34 of which were zoonotic (Fig. 4.1b). Mapping of the new EID
events followed that above for the previously gathered events. Point localities were all mapped were
exact coordinates were given in the source and the exact coordinates were found using Google Earth
v6.2. if only textual geospatial information was given. The locality of weighted centroids was assigned
to EID events with only sub-national unit and EIDs with only country localities were excluded. In total,
30 new EID events were retained.
The majority of these new EID event pathogens are viruses, with several bacteria and only two
protozoa (Table 1). Three events were due to drug-resistant strains, and only one due to a newly
virulent strain. Although data was collected with a perspective to continuing from where Jones et al.
(2008) left off, 15 EID events occurred pre-2004, 7 of which are from the 1980s or 1990s. This likely
includes previously unidentifiable outbreaks that have only been classified using modern techniques.
17 of the 30 EID event pathogens in Table 1 have identified reservoir hosts or at least known animal
infections (11 wild, 2 domestic, 4 both host types). However, most are recently discovered, therefore
natural hosts for the remaining 13 will no doubt take time to locate and confirm.
4.4 Maps of zoonotic emerging disease events
To remap the combined previous and the new EID events, we created a one decimal degree grid
using HawthsTools 3.27, and used ArcMap 9.3’s ‘Spatial Join’ tool to assign each EID event point to a
single grid cell. New zoonotic EID events were mapped and labelled (Fig. 4.2) (n=30). Although the
events are globally spread across every continent, there may be potential clusters of new or newly
identified EID events in the Northeast US, South America, Continental Europe, and Southeast Asia.
Only one grid cell contained more than one event, which was in Cameroon, containing Human T-
lymphotrophic viruses 3 and 4, which were discovered simultaneously (Table 1). Previous EID events
were mapped in combination with the new events (n=202) (Fig. 4.3). Again, events are present across
most of the inhabited world. In contrast to the new events, the clearest potential hotspots are now the
USA, and Western Europe, which likely reflects historical differences in surveillance and reporting. As
in Jones et al., the maximum number of events in any grid cell was 6, occurring in Central London,
UK. Separate maps were also produced for those events with wild hosts, and non-wild hosts (Fig. 4.4
4.5). Similar patterns were present for each, with the most noticeable difference being that very few
EID events from non-wild hosts occurred in Africa or South America. Maps were also produced to
illustrate breakdown of events in each grid cell by zoonotic host categorisation, drug resistance, and
type of data (see Figure 3).
Figure 4.1. Tree diagrams illustrating data structure for EID events from 1) Jones et al. 2008,
and 2) this update, with respect to zoonotic status, spatial data quality of entries, and whether data
was accepted or rejected for the new maps contained in this report.
Figure 4.2 New zoonotic disease events identified in 2012 and not previously mapped
Figure 4.3 Previous zoonotic emerging disease events were mapped in combination with the new events
Figure 4.4 Zoonotic emerging disease events with wildlife hosts
Figure 4.5 Zoonotic emerging disease events with non-wildlife hosts
Figure 4.6 Maps of all zoonotic emerging infectious disease events (n = 202) stratified by potential variables of interest. Size of circles denotes number of
events in each one degree grid cell, and colour denotes breakdown of events in terms of type of zoonotic host
Figure 4.7 Maps of all zoonotic emerging infectious disease events (n = 202) stratified by potential variables of interest. Size of circles denotes number of
events in each one degree grid cell, and colour denotes breakdown of events in terms of drug-resistant events
Figure 4.8 Maps of all zoonotic emerging infectious disease events (n = 202) stratified by potential variables of interest. Size of circles denotes number of
events in each one degree grid cell, and colour denotes breakdown of events in terms of spatial data resolution (see also Figure 1).
4.5 Conclusion on zoonotic emerging infectious disease vents
We have geo-located and mapped a total of 202 zoonotic emerging infectious disease events, 30 of
which were newly collected during this contract. High-resolution maps are provided for all events, as
well as events stratified by type of zoonotic host, and other potential classifiers of interest. Potential
hotspot trends could broadly reflect surveillance differences, although the higher representation of
developing countries within the new events may suggest increasing research focus or improving
diagnostic technology
(Note: Is the world becoming sicker or are we just better able to detect disease? The last decades
have seen dramatic improvements in biological disease detection with dozens of new potential
pathogens anticipated by 2020. At the same time innovations in information management are
increasing awareness of disease outbreaks. Perry et al. (2011) explore this in a recent review and
conclude that there is evidence for increased emergence of disease in recent decades, although this
is not historically unprecedented: major epidemiological transitions also occurred during the Neolithic
when livestock were domesticated on a wide-scale, during the age of exploration when Old World
pathogens were introduced to the New World, and to a lesser extent with increased global travel in the
nineteenth century).
We also updated the regional maps based on the national poverty data showing:
Numbers of cattle, sheep, goats and pigs
Human population
Poor livestock keepers
The maps and tables are shown in the this section as well as a summary of regional characteristics
relevant to the review.
Figure 5.1 Poor livestock keepers (million) by region
SouthAsia WestCentral
EastAfrica SouthEast
East Africa
Countries: Sudan, Ethiopia, Eritrea, Djibouti, Somalia, Kenya, Uganda, Rwanda, Burundi, Tanzania
Fig 5.2: Eastern Africa region: Farming systems (Herrero et al 2009, Notenbaert et al 2009)
Table 5.1: Species population by farming system in Eastern Africa (Herrero et al 2009)
Others 
Bovines32,239,10059,221,300 13,481,300 4,479,000109,420,700
Goats35,603,00031,020,500 6,699,100 3,082,78076,405,380
Sheep34,404,70029,999,200 4,893,750 2,254,03071,551,680
Pigs85,169433,390 543,184 183,7001,245,443
People30,608,70076,115,200 37,649,000 14,669,900159,042,800
12,125,10031,719,400 11,281,600 2,740,62057,866,720
East Africa is characterised by:
Poor livestock keepers 36% of the population; cattle are relatively important.
Mixed extensive systems predominate but agro-pastoral /pastoral are more important than most
other regions. Pastoralists have high vulnerability to zoonoses.
Zoonoses with a wildlife interface are important.
Rapidly dairy development in highlands: bringing risks of brucellosis, tuberculosis and milk-borne
Rapid growth in pig production in Uganda brings risks of emerging disease such as Ebola.
Intra-regional trade important for the horn of Africa (shoats), and of interest to Ethiopia.
High zoonoses burden in Ethiopia and Tanzania.
Insecurity in Somalia and possibly South Sudan with implications for zoonoses.
Southern Africa Region
Countries: Zambia, Malawi, Mozambique, Zimbabwe, Botswana, Swaziland, Lesotho, South Africa
Fig 5.3: Southern Africa region: Farming systems (Herrero et al 2009)
Table 5.2: Species population by farming system in SA (Herrero et al 2009)
Bovines9,611,3208,532,650 2,307,420 1,746,69022,198,080
Goats7,300,5106,164,890 643,834 1,301,33015,410,564
Sheep9,611,3208,532,650 2,307,420 1,746,69022,198,080
Pigs696,2861,300,700 840,703 215,2233,052,912
People17,757,10032,350,900 5,730,630 16,532,10072,370,730
6,286,60010,182,200 783,379 1,650,57018,902,749
Southern Africa is characterised by:
Poor livestock keepers 26% of population. Lowest in SSA; Cattle relatively important
Agro-pastoral /pastoral more important. Pastoralists have high vulnerability to zoonoses.
Zoonoses with a wildlife interface are important. Current interest
Significant commercial ranching and farming, with better animal health and zoonoses control.
Some countries with export potential
Better animal health services and disease reporting systems than most other SSA regions
Wide regional variation in farming systems, zoonoses and response capacity
West Africa region
Countries: Mauritania, Mali, Niger, Chad, Senegal, The Gambia, Guinea Bissao, Guinea, Sierra
Leone, Liberia, Ivory Coast, Burkina Faso, Ghana, Togo, Benin, Nigeria, Cameroon
Fig 5.4: West Africa region: Farming systems (Herrero et al 2009)
Table 5.3: Species population by farming system in WA (Herrero et al 2009)
Others Total
Bovines13,458,000 28,220,300 2,403,490 1,034,53045,116,320
Goats19,936,400 37,553,900 16,816,900 2,603,10076,910,300
Sheep13,841,500 29,747,800 8,505,920 2,024,59054,119,810
Pigs810,597 2,598,830 2,518,430 1,367,5507,295,407
People18,540,400 80,711,100 72,559,700 13,249,700185,060,900
11,343,700 38,161,100 19,231,600 2,011,42070,747,820
West Africa is characterised by:
Poor livestock keepers 38% of the population; highest in SSA. Goats (followed by sheep) relatively
most important species
Mixed extensive more important, but mixed intensifying more important than other regions of SSA.
Pastoralism/agro-pastoralism mainly in the Sahel.
Cattle important for traction, but tsetse a major barrier. Trypanotolerant cattle important in the
some areas. Little dairying and reliance on imported dairy products. Traditional dairying in the
Sahel has high risk of milk-borne zoonoses because of cultural practices.
Also high imports of poultry. Slow growth in livestock production.
High zoonoses burden in Nigeria and sub-humid coastal countries.
Insecurity in several countries
North Africa region
Countries: Western Sahara, Morocco, Algeria, Tunisia, Libya, Egypt
Fig 5.5: North Africa Region: Farming systems (Herrero et al 2009)
Table 5.4: Species population by farming system in NA (Herrero et al 2009)