Title: INFECTIOUS DISEASES AND HUMAN POPULATION HISTORY
, By: Dobson, Andrew P., Carper, E.
Robin, Bioscience, 00063568, Feb96, Vol. 46, Issue 2
Database: Academic Search Premier
INFECTIOUS DISEASES AND HUMAN POPULATION HISTORY
Contents
Parasite life-history strategies
Basic reproductive ratio, R0
Thresholds for establishment, HT
Parasites and pathogens of early human populations
The first cities
Diseases in the New World
Meanwhile, back in Europe
Spatial aggregation and average age at infection
Age-dependent mortality
Vaccination and the control of human disease
Selection for increased disease resistance
Vaccination
Infectious diseases today
Acknowledgments
References cited
Throughout history the establishment of disease has been a
side effect of the growth of civilization
From the plagues of biblical times to the HIV pandemic of
today, infectious diseases have played an indisputably major role in human
history. The continual expansion of human populations since prehistoric times
has led to successive invasions of the human population by increasing numbers
of different pathogens. Today many people's worries about emerging pathogens
have been sharply focused by the Ebola virus outbreak in Kikwit, Zaire, and by
Lyme disease and hantavirus outbreaks throughout the United States (Garrett
1994, Levins et al. 1994). In this article, we examine the infectious diseases
of humans from an ecological perspective.
Understanding pathogens at the population level is as
important in disease prevention and control as understanding pathogens at the
microscopic or molecular level. Three ecological processes are crucial in
determining the impact, persistence, and spread of pathogens and parasites: the
size and spatial distribution of the host population, the movement of infected
and susceptible hosts and vectors, and the nutritional status of the human host
population. Although medical advances continue to reduce the impact of
degenerative and self-inflicted diseases on people who can afford to pay for
treatment, the control and prevention of infectious diseases is likely to be
increasingly dependent on a solid understanding of the ecology of pathogen
transmission and persistence.
Parasite life-history strategies
The enormous array of pathogens that infect human and other
animals may be conveniently divided on epidemiological grounds into
microparasites and macroparasites (Anderson and May 1979, May and Anderson
1979). Microparasites include the viruses, bacteria, protozoa, and fungi; they
are characterized by their ability to reproduce directly within individual
hosts, their small size, their relatively short duration of infection, and the
production of an immune response in infected and recovered individuals.
Mathematical models examining the dynamics of these pathogens divide the host
population into susceptible, infected, and recovered classes. In contrast,
macroparasites (the parasitic worms, ticks, and fleas) do not multiply directly
within an infected individual but instead produce infective stages that usually
pass out of the host before transmission to another host. Macroparasites tend
to produce a limited immune response in infected hosts, are relatively
long-lived, and are usually visible to the naked eye. Mathematical models of
the population dynamics of macroparasites have to consider the statistical
distribution of parasites within the host population.
The complexities of parasite-host population dynamics may be
reduced by the derivation of expressions that describe the most important
epidemiological features of a parasite's life cycle (Anderson and May 1979,
1991, Dobson 1988, May and Anderson 1979). Three parameters are important in
describing the dynamics of a pathogen: the basic reproductive ratio (which
determines the rate at which the pathogen is likely to spread in the
population), the threshold number of hosts required for the parasite to become
established, and the mean levels of infection of the parasite in the host
population.
Basic reproductive ratio, R0
The number of individuals that each infected individual
infects at the beginning of an epidemic is formally termed R0, the basic
reproductive ratio of the disease. The first derivation of an expression for R0
was formulated by epidemiologist George MacDonald in his seminal studies of
malaria (MacDonald 1956). This work provided the key insight that if an
infection is to persist in a host population, then each infected individual
must on average transmit the infection to at least one other individual. If
this level of transmission does not occur, the infection is likely to
progressively disappear from the population.
The basic reproductive ratio for a macroparasite is defined
as the number of daughter worms (or ticks) established in a host population
following the introduction of a solitary fertilized female worm (or tick). In
both the microparasite and macroparasite cases, the resultant expression for R0
usually consists of a term for the rates of parasite transmission, divided by
an expression for the rate of mortality of the parasite in each stage in the
life cycle. Increases in host population size or rates of transmission tend to
increase R0, whereas increases in sources of parasite mortality or decreases in
transmission rate tend to reduce the spread of the pathogen through the
population.
It is also possible to derive expressions for the levels of
prevalence (proportion of the hosts infected) and incidence (mean parasite
burden) of parasites in the host populations (Anderson and May 1991). In
general, changes in the parameter values that tend to increase R0 tend also to
produce increases in the proportion of hosts infected by a microparasite, as
well as increases in the prevalence and mean burden of worm macroparasites
(Anderson and May 1979, 1991, Dobson 1988, May and Anderson 1979). In particular,
increases in the size of the host population usually lead to increases in the
prevalence and incidence of the disease in the host population.
Thresholds for establishment, HT
Epidemiological theory suggests that host population density
is critical in determining whether a pathogen can become established and remain
endemic in a population (Anderson and May 1991), a notion that was articulated
as long ago as 1927 (Kermack and McKendrick 1927, 1932). An expression for the
threshold for disease establishment, HT, may be obtained by rearranging the
expression for R0 to find the population density where R0 equals unity. The
size of the host population needed to continuously sustain new infections
usually varies inversely with the transmission efficiency of the pathogen but
directly with its virulence (case mortality rate). Mathematical expressions may
be derived for micro- and macroparasites with either simple or complex life
cycles (Diekmann et al. 1990, Dietz 1993). The resultant expressions suggest
that changes in the parameters that tend to increase R0 tend to reduce HT, and
vice versa. Therefore, more virulent species require larger populations to
sustain them, while reductions in the mortality rate of transmission stages may
allow parasites to maintain infections in populations previously too small to
sustain them. Some modern estimates suggest that a population of approximately
500,000 people is needed to produce annually the 7000 susceptible individuals
needed to sustain an endemic infection of measles (Figure 1).
An important parallel can be drawn with the concept of
population viability used in conservation biology to determine whether a
population of an endangered species is likely to persist and not go extinct;
essentially 500,000 people can be considered as the minimal viable host
population required to sustain a continuous infection of measles. Obtaining
similar empirical estimates for other infectious diseases would yield crucial
insights into the role that urban aggregations play in maintaining infectious
diseases endemic in human populations. Preliminary analysis of historical data
for human diseases in Australia suggests that populations of approximately
200,000 people are required to sustain pertussis (also known as whooping cough)
and scarlet fever, while only 50,000 people are required to sustain diphtheria
(also known as croup; Cumpston 1927).
The most parsimonious way of examining when infectious
diseases became established in human populations is to examine human
epidemiological history from the perspective of the size and spatial
distribution of human populations as they began to expand in numbers and
aggregate into the first villages, towns, and cities in different parts of the
world at different historical times.
Parasites and pathogens of early human populations
The nomadic hunter-gatherer bands that were typical of human
social structure for most of humankind's 2-million-year history probably
supported a parasite fauna closely resembling that of the higher apes, from
whom humans inherited their original collection of pathogens. Parasites with
high transmission rates that induced little or no immunity were probably the
only pathogens able to establish in the groups of around 50-100 individuals
that characterized these early societies. Pathogens such as hepatitis B,
herpes, Epstein-Barr virus, and cytomegalovirus are still found in primitive
tribes in the few remaining isolated parts of the world (Black 1975). Direct
life cycle macroparasites (those that do not require vectors for transmission),
such as pinworms, Ascaris, lice, and ticks, were probably also common in
hunter-gatherer societies, as were sexually transmitted diseases, which can be
sustained in low-density host populations. Similarly, vector-transmitted
pathogens, such as malaria and yellow fever, may have become adapted to humans
at this early stage in human history. The long latent periods associated with
the development of Plasmodium vivax, probably the oldest species of human
malaria, suggest a mechanism for parasite survival under low transmission
conditions.
The more sedentary habits of the early agriculturists
probably increased the incidence of direct life cycle macroparasites, such as
the roundworm Ascaris, mainly because of the increasingly successful
transmission of the long-lived free-living stages, which increased in numbers
around more permanent dwellings. Water supplies became contaminated with toxic
bacteria and protozoa as humans became more sedentary. Archeological studies on
human communities entering the transition between hunter-gathering and
primitive pastoralism provide evidence suggesting that a more sedentary
existence tended to lead to increased child mortality and high levels of disease.
A significant characteristic of this malnutrition was a reduced diversity in
diet as humans switched from a mixture of meat, grains, and fruit to a diet
dominated by grains (Cohen and Armelagos 1984). This change in diet was not
entirely detrimental, because storing wheat, barley, and millet in earthenware
pots may have created ideal conditions for Streptomycetes to develop. The
tetracycline-producing Streptomycetes are effective against bacteria,
rickettsiae, spirochetes, and some viruses (Bassett et al. 1980). Thus some
ancient populations would have inadvertently dosed themselves with the first
antibiotics during roughly the same historical period that they began to
acquire their first significant bacterial and protozoan pathogens.
The first cities
The first significant aggregations of humans into sedentary
agricultural settlements occurred around 5500 B.C. at Khuzistan in the valleys
of the Tigris and Euphrates in what is now Iran (Chandler and Fox 1974).
Irrigation allowed these first cities to grow; nevertheless, it took 2000 years
for these early villages to develop into the city of Uruk (which had around
50,000 inhabitants at the peak of its influence) and the surrounding towns of
Ur, Kish, Lagash, and Umma (each of which had between 10,000 and 20,000
inhabitants). The early settlements that appeared in the Indus River valley in
India at this time and in China, Egypt, and Mesoamerica consisted more of towns
and villages than cities. None of these conurbations were large enough to
continuously sustain any of today's common childhood diseases.
Only when communication between small groups of neighboring
towns began to be established is it likely that human populations became large
enough to sustain direct life cycle bacterial and viral infections. It is in
these first cities that the now common diseases of humans started to appear.
Many of the first pathogens to infect humans evolved from diseases of domestic
animals. Measles, for example, is closely related to two other
morbilliviruses--canine distemper and rinderpest (a disease of cattle)--whereas
smallpox probably evolved from cowpox. India is probably the original home of
smallpox; traditions concerning this disease have long existed amongst the
Brahmans. Rubella, typhoid, and dysentery also arose during this period.
Although written descriptions of many early disease outbreaks are almost
impossible to decipher, rabies is known from writings in Babylonian tombs from
approximately the twenty-third century B.C., while smallpox and tuberculosis were
clearly described in Chinese writings from approximately A.D. 1000.
Hippocrates (460-377B.C.) was probably the first person to
record diseases with enough precision for them to be identified today as
malaria, mumps, diphtheria, tuberculosis, and perhaps influenza. The Greeks
were definitely aware of pulmonary tuberculosis--an epidemic of tuberculosis
seems to have raged during Hippocrates' lifetime. Hippocrates also described
many other recognizable infections, including malaria, so it is apparent that these
diseases were present in Hellenic times. Interestingly, none of Hippocrates's
records indicate the presence of smallpox, measles, or bubonic plague in
ancient Greece, which suggests that the size of ancient Greek and Egyptian
cities may not have been large enough to continuously sustain measles and
smallpox infections, and these pathogens may have died out.
It is likely that many of the plagues mentioned in the Old
Testament arose when human populations became sufficiently aggregated to
sustain epidemics of new pathogens. Because human populations had not
previously been exposed to pathogens such as smallpox, rubella, and measles,
levels of immunological resistance were low and mortality rates were high.
After several visitations by a specific pathogen, levels of genetic resistance
and herd immunity start to increase, leading to reduced pathogenicity and a
more regular pattern of prevalence.
There is circumstantial evidence that the plague of Athens
(430-429B.C.), which killed 25% of the city's population, may have been measles
(McNeill 1976); contemporary descriptions of the pathology closely resemble
those from the first epidemic of measles in Hawaii. The population of Athens
was only around 155,000 at the time of the plague; therefore, if the plague was
an early outbreak of measles, the small size of the human population would have
caused the disease to die out once the initial pool of susceptible individuals
had been exhausted. In a smallpox epidemic in Rome in A.D. 165-180, 25% of the
populace died, while 5000 a day died in a second epidemic in A.D. 251-266
(McNeill 1976). These and other epidemics may have played a significant role in
strengthening the position of the growing Christian sect, whose members
believed in assisting the sick, whom the rest of society sought to avoid. These
caregivers, through repeated low-level exposures to the virus, may have built
up their immunity to the pathogen, thus appearing protected by virtue of their
religious convictions. Perhaps this so-called protection gave the Christians an
advantage during the big Roman epidemics and convinced many nonbelievers to
join their ranks.
If one examines the distribution of city sizes over the last
3500 years, it becomes clear that cities capable of maintaining an endemic
measles infection were continuously present only from around the late
seventeenth century (Figure 2). Cities capable of continuously supporting
diphtheria (circa 50,000) have been present only since the early Christian era,
while cities capable of supporting pertussis and streptococcus (scarlet fever)
have been present only since the late Middle Ages (twelfth and thirteenth
centuries). Although some large cities were present in late Roman times, these
cities declined in the Dark Ages. Most of the cities in which today's diseases
evolved were the size of large towns or the suburban districts of today's
megacities. The data in Figure 2 suggest that sequential outbreaks of measles
in different towns may have been an important historical mechanism for ensuring
the persistence of pathogens in human populations before towns and cities were
large enough to continuously support infections.
In the case of measles, persistence would have had to occur
on time scales of the order of 10 to 18 centuries before cities were large enough
to continuously sustain infections. Consequently, a curious distortion in the
present understanding of the population biology of measles has been created.
Most studies have focused on the dynamics of measles in large cities in the
United States and Europe in the mid-twentieth century (Anderson et al. 1984,
Bolker and Grenfell 1993, Grenfell 1992, Olsen and Schaffer 1990); yet for 90%
of its history, measles has persisted by moving between ephemeral populations
of susceptibles in largely rural populations (Cliff and Haggett 1980, 1984,
1988).
May and Anderson (1990) propose a mathematical model to
explain this phenomenon as an explanation for the early period of the HIV/AIDS
epidemic. The model suggests that it may have been possible for HIV to persist for
periods of up to 100 years as a sequence of outbreaks that flared and died out
in small villages. Providing a new outbreak was initiated in another village
before the pathogen had completely died out in villages where outbreaks were
occurring, then the disease could have persisted for a long time in populations
of villages that, individually, would be too small to sustain a continuous
outbreak. It would be intriguing to know if similar spatial heterogeneities
allowed measles to persist for much of human history.
Diseases in the New World
The population of the New World probably numbered around 100
million before Columbus's arrival in the fourteenth century (McNeill 1976). In
terms of disease status it resembled the Old World around the time of the birth
of Christianity. Intestinal worms and protozoan infections prevailed, and
traces of these parasites are often found in archaeological studies of bodies
from preColombian burial sites (McNeill 1976). The apparent lack of other
pathogens in the New World may be due to differences in animal husbandry
practices (Clutton-Brock 1987). In much of South America the traditional herd
animals were llamas and alpacas. Because these animals live in small groups
high in the mountains, there was little contact with humans and smaller
populations of hosts within which diseases could become established. Important
exceptions to this case were the protozoan infections that give rise to
trypanosomiasis and its close relative leishmaniasis. These pathogens are
endemic in many species of small mammals, and Trypanosoma cruzi is particularly
prevalent in the populations of guinea pigs traditionally kept in large numbers
by New World farmers as a source of meat. Triatomid bugs transmitted the
pathogen between guinea pigs, humans, and reservoir hosts such as possums.
Disease again worked in the favor of the Christians during
the European invasion of the New World during the fifteenth and sixteenth
centuries--an event that had immense epidemiological repercussions for the
indigenous people of the Americas (McNeill 1976). The coming of the Europeans
provides perhaps the best illustration of the effects of population movement on
pathogen transmission. The introduction of smallpox, measles, and typhus to
South American and Central American human populations, who had no natural
resistance or immunity to these diseases, led to apalling numbers of deaths.
Largely as a result of successive epidemics of these diseases, the population
of Mexico fell from 20 million to approximately 3 million in the 50 years from
1518 until 1568 and then to 1.6 million in the next 50 years. The successful
colonization of these continents probably owes considerably more to the
pathogens that Europeans brought with them than to any of their more
traditional weapons.
While malaria was prevalent in Europe at the time of the
colonization of the New World, historical data suggest that the disease was
absent in the Americas before the arrival of the Europeans. Moreover, the
importation of slaves from highly endemic areas of Africa fueled epidemics that
spread in waves across the continents, ravaging both the indigenous and
colonist populations. In North America, the spread and persistence of malaria
were closely associated with pioneer life. Although epidemic levels of malaria
had subsided in the northeast colonies by the mid-1700s, later settlers in the
upper Mississippi Valley suffered high levels of mortality and morbidity. So
bad was the situation that one writer declared that "it is unlikely that
any amount of toil... will make this land habitable." The prevalence of
malaria finally began to decrease in the latter half of the nineteenth century,
probably due to improved housing, nutrition, and agricultural practices as the
frontier became settled. In the southern states, however, the disease remained
deeply entrenched as late as the 1930s and 1940s, when 4 million cases per year
were being reported. This high endemicity seeded a new round of malaria
epidemics in the northeast when Northern soldiers returned from the South at
the end of the Civil War--epidemics that did not subside until the early 1900s.
Meanwhile, back in Europe
The first influenza epidemic in Europe occurred from 1556 to
1560 (McNeill 1976), at approximately the same time as a similar epidemic in
Japan (1556). This directly transmitted microparasite has a short latent period
and is highly contagious, requiring large populations in order to establish
itself. Its first recorded appearance in the human population seems to have
caused around 20% mortality (McNeill 1976). The great influenza pandemic that
followed the World War I in 19181919 led to the deaths of 20 million
people--more than had died in the war itself (McNeill 1976). This high
mortality may have been associated with the low levels of nutrition in the
human population following the wartime rationing and the high incidence of
secondary infections that followed the war in Europe.
It is only from around 1650 that population statistics are
reliable and that real estimates can be made of the birth and survival rates of
human populations. These and similar data sets suggest that disease epidemics
were a common consequence of increasing urbanization. The data for the city of
London for the time period 1625-1750 illustrate two major outbreaks of plague
(Figure 3; Appleby 1975). Although this disease died out after thatched roofs
were replaced with slate roofs in the aftermath of the Great Fire of London,
other pathogens became established in this period. So-called fever deaths may
represent mortality due to influenza, while deaths due to consumption reflect
the increasing incidence of tuberculosis. Smallpox remained endemic in the
population at this time and caused regular epidemic outbreaks.
Spatial aggregation and average age at infection
Although diseases of childhood were well established in
towns in the seventeenth and eighteenth centuries, they were still transmitted
only intermittently in the smaller populations of rural areas. Before 1800,
less than 2% of the European population lived in cities of 100,000 or more
(McNeill 1976). Many contagious diseases were spread when boys and men joined
the army or when children went to school. Even by the time of Napoleon (1812),
it was noted that skinny and ill-fed urban recruits survived much longer than
did large, muscular, and well-fed recruits from the countryside, who were
repeatedly ill (McNeill 1976). This disparity in health was probably due to a
reduced exposure to pathogens and hence a lower level of immunity in the
smaller populations of isolated towns and villages. The earlier figure for the
size of the world's cities illustrates the aggregation of Europe's growing
human population into an increasing number of larger towns and cities (Figure
2). This large-scale change in spatial distribution was matched by a reduction
in spatial aggregation at a lower scale.
Many demographers argue that the so-called demographic
transition to smaller family sizes (Figure 4) is a direct consequence of
increased health and hence increased survival. Yet it should be noted as well
that reductions in family size also lead to direct improvements in the health
and welfare of women and perhaps to significant reductions in the rates of
disease mortality within families. Studies of mortality from measles in
families of different sizes indicate that mortality rates among children are
higher in larger families, even when the effects of increased malnutrition in
many larger families are controlled (Aaby et al. 1988, Garenne and Aaby 1990).
In general, as more children in a family succumb to an infection, its pathology
becomes worse. This outcome may be because susceptible children in larger
families receive higher disease inocula from their infectious siblings, or it
may reflect rapid selection that allows the virus to exploit the common genetic
background of children in the same family. Probably, both of these effects
operate. If they do, then the net impact of pathogens on the host population
may have declined as human societies have shifted toward smaller families.
Age-dependent mortality
The net impact of a pathogen on a population is influenced
by the interaction between the average age of infection and the age-dependent
mortality rate of the disease. If pathogen-induced mortality increases
significantly with age, then the later ages of infection that characterize
rural communities may well cause higher mortality in lower density rural
populations. In contrast, mortality may be lower in urban populations where
people are likely to be exposed to infection at earlier ages. Anecdotal
evidence of this effect is provided by information on recruits to Napoleon's
army.
A remarkable study of an outbreak of measles in an isolated
island off the northern coast of Scotland (Panum 1940) offers quantitative
evidence in support of age-dependent mortality. In 1846, when the human
population of the Faeroe Islands numbered 7782 people, measles appeared and
infected nearly everyone. A previous epidemic in 1781 had also infected almost
the whole population, but the disease had disappeared completely, presumably
due to a lack of susceptible persons. Measles was reintroduced to the
population in 1846 by an infected carpenter who had been exposed to the disease
just before leaving Copenhagen, Denmark. He became sick in April, soon after
his arrival on the islands, and by October the disease had spread throughout
the community. The 92 people who survived the 1781 epidemic escaped infection
in 1846. Others escaped by rigid quarantine and isolation, but 95% of the rest
of the population, approximately 6000 people out of a possible 6682, eventually
contracted measles. The infection then disappeared from the islands, again
presumably due to a lack of susceptible persons. The mortality rate of people
of different ages, carefully recorded by Panum (1940), reveals an interesting
age-dependent case fatality rate (Figure 5). Children younger than one year old
tended to die from measles. The mortality rate then declined to approximately
0.3% in age classes up to age 30, after which it increased first to 1.5%, and
then to 8%-10% in people older than 60. Thus the average age of infection is
likely to be important in determining the impact a disease such as measles has
on a population. If this relationship between mortality and age occurs for
other infectious diseases, then infrequent outbreaks of disease in isolated and
semi-isolated rural and island populations would lead to higher levels of
mortality than would occur in cities, where the majority of individuals are
exposed to common, so-called childhood diseases when still young.
Vaccination and the control of human disease
The migration of people into the industrial cities in Europe
throughout the nineteenth century probably increased the rates of transmission
of directly transmitted diseases. Before 1900 it is unlikely that populations
in cities were self-sustaining, because before this date urban populations were
maintained only by immigration from the surrounding countryside, where 80%-90%
of the population still lived. The mortality rates caused by common childhood
pathogens began to decline during this period, particularly when vaccines (for
smallpox) or treatments (for scarlet fever) became available (Figure 6). Notice
in Figure 6 that several of the diseases exhibit a pronounced tendency to
produce regular epidemic cycles. The period of these cycles can be roughly
estimated using a simple formula (period = 2 Pi(AD)1/2, where A is average age
of infection and D is the duration of time for which an individual is infected;
Anderson and May 1982). This relationship suggests that the interepidemic period
is likely to decline as the average age at infection declines (Figure 7).
Several of the data sets (measles and smallpox) illustrated in Figure 6 show an
apparent shortening of interepidemic period as we move from the mid-nineteenth
to the early twentieth century. This shorter period may reflect the younger age
at first infection (and hence increasing R0) as the human population became
larger and more aggregated into urban areas. It may also reflect the effects of
better nutrition in reducing the time that people were infectious.
Selection for increased disease resistance
There are almost no data available that can be used to
examine selection for resistance in humans to the common childhood diseases
such as measles, mumps, and rubella. Similarly, there are no studies that allow
scientists to accurately examine reductions in the virulence of these pathogens
(Ewald 1983). There is one historical study of tuberculosis (Ferguson 1933,
1934) that allows some quantification of the selection pressure that this pathogen
may have placed on previously naive human populations. In a study of
tuberculosis in American Indians of the Canadian plains, Ferguson (1933,1934)
records huge increases in mortality due to tuberculosis once the American
Indians were removed from their normal nomadic existence and settled onto
reservations (Figure 8). The mortality rate from tuberculosis, which had not
previously been recorded for these tribes, rapidly increased to 90-140 cases
per 1000 people within a few years of the reservations' founding. From 1882
through the 1930s, the mortality rate decreased from 90 cases per 1000 people
to 20 cases, and then to 8 cases, per 1000 people (Ferguson 1933, 1934). In
some years the mortality rate was 20 times the rate for the surrounding white
population. A number of different factors may have caused this high mortality
rate--inadequate diet after the disappearance of buffalo, poor housing,
overcrowding, and general spiritual demoralization. Yet many European
immigrants exposed to conditions as bad as those encountered by the American
Indians did not have mortality figures of this magnitude. Careful analysis of
the family trees of the Indians by Ferguson revealed that during the peak of
the epidemic in the 1880s, although some families suffered high mortality,
others suffered much lower mortality. Many of the surviving American Indian
families were descendants of the families with lower mortality. This finding
suggests that some selection for reduced susceptibility to tuberculosis
operated.
Unfortunately, it may never be possible to determine the
relative importance for declining disease mortality rates of changing age at
first infection, smaller family size (and hence smaller inocula), better
nutrition, and selection for resistance. Obviously improvements in nutrition
and personal hygiene have been important, but changes in the spatial
distribution of humans and selection for reduced virulence in the pathogens may
also have played important roles in changing average age of infection and case
mortality rate. Furthermore, most of the reduction in pathogen-induced
mortality rates occurred before the full development of preventive medicine,
suggesting that much of the reduction of the impact of disease may have been
due to the development of agricultural practices that made food cheaper and
more plentiful and the development of more efficient means of sewage and waste
disposal.
Vaccination
The development of vaccines for many directly transmitted
microparasitic diseases of humans contributed to the final decline in juvenile
mortality rates in developed countries during the present century.
Nevertheless, these diseases are still major causes of death in the developing
world. Mathematical population models can again help to explain why vaccination
has been successful against some pathogens and why it may be less successful
against others. Essentially, a successful vaccination scheme should reduce the
size of the pool of susceptibles to below the threshold needed for the pathogen
to sustain itself. More specifically, the proportion of the host population
that needs to be vaccinated, p, is given by p=1-1/R0 (Figure 9). It is thus
possible to eradicate a disease by vaccinating only a proportion of a
population rather than the whole population. However, this possibility also
implies that although vaccination may be useful as a control against diseases
such as rubella and measles, it is unlikely to be effective against diseases
such as malaria, where a large proportion of the infant population is infected
by the age of six months.
One of the major additional effects of vaccination is the
reduction of infected individuals in the population due to the reduced number
of contacts that infected individuals have with susceptible individuals. This
effect is called herd immunity (Fine 1993, Fox 1983, Fox et al. 1971). As the
percentage of the population that is immunized increases, there are linear
decreases in the total incidence of the disease in the population and
increasingly rapid decreases amongst the proportion of individuals not
vaccinated.
Arita et al. (1986) assembled data on population densities
for smallpox vaccination coverage in African and Asian countries during the
late 1960s and early 1970s. These data indicate that smallpox disappeared early
from countries in which the density of susceptible (unvaccinated) individuals
fell below ten persons per square kilometer. In populations with a population
density of less than 50 persons per square kilometer, this density roughly
corresponds to 80% coverage. Infections of smallpox persisted in more densely
populated regions, in particular Nigeria, with 58 persons infected per square
kilometer; Pakistan, with 83 persons per square kilometer; India, with 175
persons per square kilometer; and Bangladesh, with 502 persons per square
kilometer (Arita et al. 1986). To eradicate smallpox in these areas would have
required coverage of the order of 98% to reduce the number of susceptible
individuals to less than ten persons per square kilometer; such coverage was
impractical. By 1970 vaccination policy changed to active case detection,
contact tracing, and the breaking of individual chains of transmission. It was
this policy change that eventually led to the worldwide eradication of smallpox
(Arita et al. 1986, Fine 1993).
Infectious diseases today
Despite widespread advances in medical science, infectious
diseases continue to have devastating consequences for human populations in
many parts of the world. Although the incidence of measles in the United States
has fallen by approximately 99% since the introduction of vaccination in 1963,
annual resurgences continue even though approximately 98% of children in the
United States have been vaccinated by the age of school entry. Analysis of
surveillance data suggests that transmission has been continuous in several
large urban populations, in particular those with large, poor, inner-city
populations--such as New York, Los Angeles, and Chicago. Although transmission
is only sporadic throughout the rest of the country, inner-city areas present
an extremely difficult problem for public health providers because the social
conditions for high vaccine coverage are least conducive in the areas where the
highest uptake is required. As urbanization trends continue globally, disease
is likely to continue to be maintained in cities, presenting a constant threat
to the decreasing proportion of the human population living in rural areas.
In 1993 alone, tuberculosis killed 2.7 million people and
infected another 8.1 million worldwide (Snider 1994). An estimated one-third of
the world's population, or 1.7 billion people, is infected but has not yet
developed the disease. Furthermore, the mortality rates of infected individuals
are sharply related to socioeconomic status (Figure 10).
The present tuberculosis epidemic is expected to grow worse,
especially in the developing world. This epidemic is partly driven by the
increasing urban population, by the evolution of multidrug-resistant strains,
and by the emergence of HIV/AIDS, which compromises the immune system of human
hosts and makes them more susceptible to infectious diseases such as
tuberculosis. In the United States and Europe, prisons and homeless populations
now act as reservoirs for the resurgence of tuberculosis. These underserved
subpopulations are now above the threshold sizes for maintaining diseases in
areas where they had previously been significantly reduced.
While medical and sanitation technology have certainly
helped minimize the impact of infectious diseases in the last two centuries,
humans can only really claim victory over one pathogen--the smallpox virus. It
seems inevitable that other diseases (e.g., measles and typhus) that may now be
simmering quietly in scattered populations will reemerge as the susceptible
subpopulations of major cities increase to threshold sizes.
The epidemiological and nutritional situation of many
developing countries is similar in some ways to that of Europe in the
seventeenth and eighteenth centuries (Livi-Bacci 1991). Although smallpox has
been successfully eradicated, malaria, measles, rubella, and other childhood
diseases are still major causes of death. Malaria, in particular, is showing a
major resurgence. Following the widespread use of drugs and insecticides
against the pathogen and its mosquito vector in the mid-twentieth century,
genetic resistance has evolved to these compounds, and malaria is returning
rapidly to areas that had been declared free of the infection. The
macroparasitic pathogens such as hookworm, ascaris, and schistosomiasis
continue to debilitate many millions of people. The effects of all diseases in
the developing nations are made worse by poor nutrition. The United Nations
Children's Fund estimates that 50%-75% of all recorded deaths of infants and
young children can be attributed to a combination of malnutrition and infection
(UNICEF 1995).
As the only developed nation other than South Africa without
a national health plan, the United States may be the first developed country to
have to contend with reemerging diseases. For example, although malaria had
been successfully controlled in most of the country by the mid-nineteenth
century, large cities such as New York still supported endemic transmission of
Plasmodium falciparum as late as the early 1940s. In an interesting parallel to
current HIV epidemiology, falciparum malaria was maintained in the
mid-twentieth century entirely among intravenous drug users (Blower et al.
1991, Most 1940). Today, with access to quality health care dependent on
socioeconomic status, the rapid growth of susceptible subpopulations (e.g.,
homeless and prison populations) in large cities is not surprising. These
subpopulations are susceptible not because they have no history of exposure to
a particular pathogen (the majority of people in developed countries have not
been exposed to many infectious pathogens) but because they are underserved by
or inaccessible to the medical community (Berkelman et al. 1994).
No challenge to medical science more sharply illustrates the
importance of developing an ecological understanding of infectious disease than
the current pandemic of HIV/ AIDS. Much of the progress in understanding
HIV/AIDS comes from studies at the population level (Anderson and May 1991) and
from extensions of these mathematical techniques to develop an understanding of
the population dynamics of interactions between components of the immune system
and the HIV virus in individual infected hosts (Nowak et al. 199]). These
quantitative scientific insights into the biology and epidemiology of HIV have
been obtained using extensions of models originally developed to study the
ecology of infectious diseases in human and other animal populations (Anderson
and May 1991). It is essential that policy for disease-control strategies in
the twenty-first century build upon this understanding of diseases at the
population level as well as at the molecular and cellular level.
Acknowledgments
We would like to thank Leslie Real for comments on an
earlier draft and for inviting us to contribute to this special issue. We would
also like to thank Bryan Grenfell, Bob May, Annarie Lyles, Ruth Berkelmen, and
the students of Freshman Seminar 102 "AIDS, Anthrax and Worms: Disease in
Human History" for many insightful and stimulating discussions. We also
thank David Goodman and Louise Schaeffer of the Biology Library at Princeton
University for never failing to find the obscure books we request. Andrew P.
Dobson's work is partly sponsored by the Geraldine R. Dodge Foundation.
GRAPH: Figure 1. The relationship between human population
size on oceanic islands and the percentage of months when measles was recorded
on the islands (after Black 1966).
GRAPH: Figure 2. Distribution of population size of the
world's cities from 1360 B.C. until A.D. 1968. The data are divided into
frequency classes that increase on a logarithmic scale (e.g., 1000, 2000, 5000,
10,000). Unfortunately, it is logistically impossible to obtain information on
historic aggregations of less than 2000-5000 people, or more recent data for
towns of less than 20,000 people (data from Chandler and Fox 1974).
GRAPHS: Figure 3. Recorded deaths from different diseases in
the city of London 1625-1750. (a) During the period 1625-1700, plague had a
dramatic impact on the population. (b) and (c) During the period 1650-1750,
there were cyclic outbreaks of smallpox (10003000 cases per year; lower lines)
and a steady rise in deaths from consumption (tuberculosis [TB]; upper lines).
Redrawn from data in Appleby 1975.
GRAPH: Figure 4. Frequency distribution of family sizes in
Great Britain 1870-1925 (after Wrigley 1969). Decreases in family size may have
lead to changes in the average age at infection and the case mortality rate.
GRAPH: Figure 5. The age-dependent mortality rate due to an
outbreak of measles in the population of the Faeroe Islands in 1846 (after
Panum 1940).
GRAPHS: Figure 6. Decline in mortality rates from four
common microparasitic diseases in England and Wales during the second half of
the nineteenth century: (a) measles, (b) pertussis, (c) scarlet fever
(Streptococcus), and (d) smallpox. In each case the figures given are death
rate per million people infected (after Cumpston 1927).
GRAPH: Figure 7. The relationship between average age of
infection, A, and interepidemic period for three different periods for which
the host is infected: one week (bottom line), two weeks (middle line) and four
weeks (top line). The period of infection combines the incubation (or latent)
and infectious periods of the pathogen. The values of R0 corresponding to this
particular age of infection are given above the main part of the graph; this
calculation assumes life expectancy (L) = 50 years. Typical periods of
infection for some common childhood diseases are: measles (12-16days), mumps
(16-24days), pertussis (28-33days), rubella (18-26days), smallpox (10-15days),
and scarlet fever (15-23days). These periods may be longer in malnourished or
immunocompromised hosts.
GRAPH: Figure 8. The mortality rate due to tuberculosis in
three generations of Canadian plains American Indians after their forced
adoption of a sedentary existence (after Ferguson 1933).
GRAPH: Figure 9. The relationship between the proportion of
a community that must be immunized to eradicate an infection, p, and the basic
reproductive rate, R0. The figure also illustrates the average of infection, A,
before vaccination for different values of R0; this relationship is shown for L
= 70 (lower figures) and L = 45 (upper figures). The infection is eradicated
for values of p above the line.
GRAPH: Figure 10. The incidence of tuberculosis in different
socioeconomic sectors of the population in New York State in 1976 (Hinman et
al. 1976).
References cited
Aaby P, Bukh J, Dasilva MC, Lisse IM. 1988. Decline in
measles mortality--nutrition, age at infection, or exposure. British Medical
Journal 296: 1225-1228.
Anderson RM, May RM. 1979. Population biology of infectious
diseases: Part I. Nature 280: 361-367.
------. 1982. Directly transmitted infectious diseases:
control by vaccination. Science 215: 1053-1060.
------. 1991. Infectious diseases of humans: dynamics and
control. Oxford (UK): Oxford University Press.
Anderson RM, Grenfell BT, May RM. 1984. Oscillatory
fluctuations in the incidence of infectious disease and the impact of
vaccination: time series analysis. Journal of Hygiene 93: 587-608.
Appleby AB. 1975. Nutrition and disease: the case of London,
1550-1750. Journal of Interdisciplinary History 6: 1-22.
Arita I, Wickett J, Fenner F. 1986. Impact of population
density on immunization programmes. Journal of Hygiene 96: 459-466.
Bassett EJ, Keith MS, Armelagos GJ, Martin DL, Villanueva
AR. 1980. Tetracycline-labeled human bone from ancient Sudanese Nubia (A.D.
350). Science 209: 1532-1534.
Berkelman RL, Bryan RT, Osterholm MT, LeDuc JW, Hughes JM.
1994. Infectious disease surveillance: a crumbling foundation. Science 264:
368-370.
Black FL. 1966. Measles endemicity in insular populations:
critical community size and its evolutionary implication. Journal of
Theoretical Biology 11: 207-211.
------. 1975. Infectious diseases in primitive societies.
Science 187: 515-518.
Blower SM, Hartel D, Dowlatabadi H, Anderson RM, May RM.
1991. Drugs, sex and HIV: a mathematical model for New York City. Philosophical
Transations of the Royal Society of London B Biological Sciences 321: 171-187.
Bolker B, Grenfell BT. 1993. Chaos and biological complexity
in measles dynamics. Proceedings of the Royal Society of London B Biological
Sciences 251: 75-81.
Chandler T, Fox G. 1974. 3000 years of urban growth. New
York: Academic Press.
Cliff AD, Haggett P. 1980. Changes in the seasonal incidence
of measles in Iceland, 1896-1974. Journal of Hygience 85: 451-457.
------. 1984. Island epidemics. Scientific American 250: 38.
------. 1988. Atlas of disease distributions: analytic
approaches to epidemiologic data. Oxford (UK): Basil Blackwell.
Clutton-Brock J. 1987. A natural history of domesticated
mammals. Austin (TX): University of Texas Press.
Cohen MN, Armelagos GJ. 1984. Paleopathology at the origins
of agriculture. New York: Academic Press.
Cumpston JHL. 1927. The history of diptheria, scarlet fever,
measles, and whooping cough in Australia, 1788-1925. Canberra (AUT):
Commonwealth of Australia, Department of Health.
Diekmann O, Metz JAJ, Heesterbeck H. 1990. On the definition
and the computation of the basic reproductive ratio R0 in models for infectious
diseases in heterogeneous populations. Journal of Mathematical Biology 28:
365-382.
Dietz K. 1993. The estimation of the basic reproductive
number for infectious diseases. Statistical Methods in Medical Research 2:
23-41.
Dobson AP. 1988. The population biology of parasite-induced
changes in host behavior. Quarterly Review of Biology 63: 139-165.
Ewald PW. 1983. Host-parasite relations, vectors, and the
evolution of disease severity. Annual Review of Ecology and Systematics 14:
465-485.
Ferguson RG. 1933. The Indian tuberculosis problem and some
preventative measures. National Tuberculosis Association Transactions 29:
93-106.
------. 1934. Some light thrown on infection, resistance and
segregation by a study of tuberculosis among Indians. Transactions of the
American Clinical and Climatological Association 20: 18-26.
Fine IDEM. 1993. Herd immunity: history, theory, paractice.
Epidemiologic Reviews 15: 265-302.
Fox JP. 1983. Herd immunity and measles. Reviews of
Infectious Diseases 5: 463-466.
Fox JP, Elveback L, Scott W. 1971. Herd immunity: basic
concept and relevance to public health immunization practices. American Journal
of Epidemiology 94: 179-189.
Garenne M, Aaby P. 1990. Pattern of exposure and measles
mortality in Senegal. Journal of Infectious Diseases 161: 1088-1094.
Garrett L. 1994. The coming plague: newly emerging diseases
in a world out of balance. New York: Farrar, Straus & Giroux.
Grenfell BT. 1992. Chance and chaos in measles dynamics.
Journal of the Royal Stastical Society B. 54: 383-398.
Hinman AR, Judd JM, Kolnik JP, Daitch PB. 1976. Changing
risks of tuberculosis. American Journal of Epidemiology 103:486-497.
Kermack WO, McKendrick AG. 1927. A contribution to the
mathematical theory of epidemics. Proceedings of the Royal Society of London A
115: 700-721.
------. 1932. A contribution to the mathematical theory of
epidemics. Part II. The problem of endemicity. Proceedings of the Royal Society
of London A 138: 55-83.
Levins R, Awerbuch T, Brinkman U, Eckardt I, Epstein P,
Nakhoul N, Albuquerque de Possas C, Puccia C, Spielman A, Wilson MA. 1994. The
emergence of new diseases. American Scientist 82: 52-60.
Livi-Bacci M. 1991. Population and nutrition: an essay on
European demographic history. Cambridge (UK): Cambridge University Press.
MacDonald G. 1956. Theory of the eradication of malaria.
Bulletin of the World Health Organization 15: 369-387.
May RM, Anderson RM. 1979. Population biology of infectious
diseases: Part II. Nature 280: 455-461.
------. 1990. Parasite-host coevolution. Parasitology 100:
S89-S101.
McNeill WH. 1976. Plagues and people. New York: Anchor
Press/Doubleday.
Most H. 1940. Falciparum malaria in drug addicts. American
Journal of Tropical Medicine and Hygiene 20: 551-569.
Nowak MA, Anderson RM, McLean AR, Wolfs TFW, Goudsmit J, May
RM. 1991. Antigenic diversity thresholds and the development of AIDS. Science
254: 963-969.
Olsen LF, Schaffer WM. 1990. Chaos versus noisy periodicity:
alternative hypotheses for childhood epidemics. Science 249: 499-504.
Panum PL. 1940. Observations made during the epidemic of
measles on the Faeroe Islands in the year 1846. Delta Omega Society.
Snider DE. 1944. Tuberculosis: the world situation. History
of the disease and efforts to combat it. Pages 13-33 in Porter JDH, McAdam
KPWJ, eds. Tuberculosis: back to the future. New York: John Wiley & Sons.
Wrigley EA. 1969. Population and history. New York: McGraw
Hill.
[UNICEF] United Nations Children's Fund. 1995. State of the
world's children 1995. Oxford (UK): Oxford University Press.
~~~~~~~~
By Andrew P. Dobson and E. Robin Carper
Andrew P. Dobson is an assistant professor in the Department
of Ecology and Evolutionary Biology,