Defining chronic diseases
• Chronic diseases are difficult to define by using
the well-known criteria of:
–
–
–
–
–
–
causation,
acuteness,
age of onset,
activity restriction,
period of illness and
premature mortality.
1
Defining chronic diseases …
• They are mostly characterized by:
–
–
–
–
–
complex causality,
multiple risk factors,
a long latency period,
a prolonged course of illness, and
functional impairment or disability.
2
Arrival of the chronic disease epidemic
• Chronic diseases are not new to human societies.
Prehistorical and historical evidence clearly shows
that our ancestors suffered from a variety of
chronic diseases including osteoarthritis and
diabetes (Hinkle 1987).
• However, it was not until the twentieth century
that chronic diseases began to dominate the health
scene; their prevalence is now greater than at any
earlier period in human history
3
Arrival of the chronic disease epidemic …
• Up until the nineteenth century, infectious diseases
and injury dominated the health of all populations
even though descriptions of chronic diseases and
conditions such as coronary heart disease,
atherosclerosis and diabetes among affluent
sections of the society were beginning to
accumulate (Cohen 1989).
• The growing impact of chronic diseases and
conditions such as stroke, cancer, renal disease
and high blood pressure was not fully appreciated
until the early part of the twentieth century.
4
Geography of Chronic Disease
• What are the determinants/risk factors of chronic
diseases such as heart disease, cancer, asthma?
• Why a geography of chronic disease?
5
Why a Geography of Chronic Disease? …
• Examining chronic disease at the lowest available
level of data aggregation (local level) is beneficial
in highlighting localized rate variations.
• Geographic patterns that remain after variations
due to known factors have been extracted can be
used to suggest new factors responsible for the
remaining variations in measures of chronic
disease e.g., cancer
• Such information enable public health officials to
target additional resources to specific locations.
6
Coronary Heart Disease (CHD)
• One in five men and women have some form of
CVD.
• Although the odds for men to develop a CVD
before the age of 60 exceed that of women (one in
three for men and 1 in 10 for women), women
develop CVD at a higher rate in the
postmenopausal years.
7
Coronary Heart Disease (CHD)
• Coronary heart disease (CHD) mortality is
associated with social and material deprivation,
though the principle feature of sudden and
unexpected coronary heart disease mortality, is
uncertain.
• In general, instances of mortality from CHD is
not uniformly distributed throughout regions, as
distinct geographical ‘hot spots’ exist where
mortality is usually greater than expected.
8
Examples of Geographical studies of CVD
• Soo et al. (2001 – Nottinghamshire, England)
• Barker and Martyn, (1991 – England & Wales)
9
Geography of Coronary Heart Disease
Nottinghamshire, England
• Soo et al. 2001 conducted a population based,
retrospective study in the County of Nottinghamshire,
England
• County has 191 electoral areas. In the 4 years from 1
January, 1991 to 31 December, 1994
• 1634 patients sustained a cardiac arrest attributed to a
cardiac cause
• The overall crude mean incidence rate of community
cardiac arrest per electoral area was 40.2 per 100 000
10
population (range 0–121.2).
Geography of Coronary Heart Disease
Nottinghamshire, England …
• Thirteen electoral areas, relatively deprived according to
the Townsend score, had a significantly greater than
expected incidence rate of cardiac arrest (median of
75.6:100 000 per electoral area; interquartile range (IQR)
65.3, 83.8).
• Twelve relatively affluent electoral areas had a
significantly lower than expected incidence rate (median
of 18.5:100 000 per area (IQR 13.0, 28.7).
• See map below.
11
Geography of Coronary Heart Disease
Nottinghamshire, England …
12
Geography of Coronary Heart Disease
Nottinghamshire, England …
• After adjusting for deprivation index, there
were no differences in coronary heart disease
(CHD) mortality and community cardiac arrest
in urban and rural electoral areas.
• Apart from response times by ambulance crews,
the events that follow the cardiac arrest such as
bystander resuscitation, ventricular fibrillation
found as the presenting rhythm and survival were
similar in all electoral areas.
13
Geography of Coronary Heart Disease
Nottinghamshire, England …
• Conclusions
• Increasing level of deprivation is associated with
areas of increased incidence of out-of-hospital
cardiac arrest in Nottinghamshire, and the effect is
apparently different from that on CHD mortality.
• Strategies should aim at improving survival in areas
identified as having high rates of community cardiac
arrest.
14
Geographical studies of CVD
• Barker and Martyn, (1991 – England & Wales)
• A possible explanation for the geographical
differences in mortality from cardiovascular
disease in England and Wales is that its causes
begin to operate not in adult life but during fetal
development and infancy.
– Infant mortality
– Animal studies
– Studies in Humans
15
Infant mortality and CVD
• Infant mortality is, of course, no more than a
general indicator of adverse environmental
conditions
• Records of infant mortality dating from the
beginning of the century allow current death rates
in the 212 local authority areas of England and
Wales to be compared with infant mortality rates
in the same places 60 or more years ago.
16
Infant mortality and CVD …
• The correlation between past infant mortality
and current mortality from cardiovascular
disease (r = 0.73) is strikingly close.
• This relationship implies that some aspect of poor
living conditions in early childhood determines risk
of cardiovascular disease in adult life.
17
Infant mortality and CVD …
• The records of infant mortality in England and Wales
are sufficiently detailed to allow neonatal mortality
(i.e., deaths before one month of age) to be
distinguished from post-neonatal mortality (i.e., deaths
between the ages of one month and one year).
• A further analysis using these separate categories
showed that adult cardiovascular mortality is more
closely linked to neonatal mortality 60 years earlier
than to post-neonatal mortality
• Neonatal mortality in the past was high in places
where many babies were born with low birth weight.
18
Infant mortality and CVD …
• Neonatal mortality is also known to have been
associated with high maternal mortality.
• High rates for both neonatal and maternal mortality
were found in places where the physique and health of
women were poor.
• Cardiovascular disease is therefore associated more
strongly with poor maternal physique and health
and poor fetal growth than with conditions, such as
overcrowding, that predispose to high post-neonatal
death rates.
19
Animal studies and CVD
• Ideas about the importance of early life in
determining risk of disease in adulthood are
reinforced by studies in animals
• Transient events in prenatal or early postnatal life
have permanent and profound effects on
physiology though such effects may remain latent
until the animal is mature.
20
Animal studies and CVD …
• A female rat injected with a few micrograms of
testosterone propionate during the first four days of
life develops normally until puberty.
• Only then does it become apparent that the
hypothalamic neuronal substrate that mediates the
cyclic release of gonadotrophins has been
irreversibly altered to a male pattern when, despite
adequate ovarian and pituitary function, the animal fails
to ovulate or show normal patterns of female sexual
behaviour.
• The same injection of androgen given when the animal
is 10 days old has no effect on reproductive function.
21
Animal studies and CVD …
• In another example, the nutrition of pregnant and
lactating rats was manipulated.
• The adult body size of these rats was more powerfully
determined by their mothers' nutrition during pregnancy
and lactation than by their genetic constitutions.
• Under-nutrition during pregnancy stunted the growth of
the offspring and this effect could not be reversed by an
optimal diet after birth.
22
Animal studies and CVD …
• Nutritional deprivation in early life affects the size and
DNA content of different organ systems, depending on
the precise time at which it occurs.
• In rats, a brief period of energy restriction immediately
after birth caused a profound reduction in the weight of
the liver, spleen, and thymus, while brain and skeletal
muscle were spared.
• Energy restriction immediately after weaning reduced
only the weight of the thymus.
23
Studies in Humans and CVD
• The question explored is whether the programming
effect of the early environment applicable to the
pathogenesis of cardiovascular disease in humans
• This was done by studying adults in middle and old
age whose growth and development in infancy was
recorded in Hertfordshire, UK.
• From 1911 onwards, every baby born in the county
of Hertfordshire was weighed at birth, visited
periodically by a health visitor throughout the first
year, and weighed again at one year of age.
24
Studies in Humans and CVD …
• The records of these visits have survived so that it
is possible to trace men and women born about 60
years ago and to relate these measurements to the
later occurrence of illness and death and to the
level of known risk factors for cardiovascular
disease.
• In the first study, 6500 men born in eight districts
of the county between 1911 and 1930 were
followed up.
25
Studies in Humans and CVD …
• Table I shows their standardized mortality ratios for
ischemic heart disease according to weight at one
year; the ratios fall steeply as weight at one year
increases. There are similar trends with increasing
birth weight, though the relation is not as strong as
with weight at one year.
• These findings prompt questions about mechanism.
• There is now evidence that haemostatic variables,
glucose tolerance, blood pressure, and lipid
metabolism are all susceptible to the programming
effects of the environment in early life.
26
Studies in Humans and CVD …
Standardized mortality ratios for IHD according to weight at one year in
6500 men during 1911-1930 in Hertfordshire, England.
Weight at one year ( pounds)
Ischemic heart
Disease (deaths)
All non-circulatory
disease
≤ 18
100(36)
74(39)
-20
84(90)
99 (157)
-22
92 (180)
74 (215)
- 24
70 (109)
67 (155)
- 26
55(44)
84(99)
≥27
34(10)
72(31)
All
78(469)
78(696)
27
Source: Barker & Martyn 1991
Studies in Humans and CVD …
• High plasma concentration of fibrinogen is
a strong predictor of increased risk of both
ischemic heart disease and stroke.
• Fibrinogen concentrations have been
measured in 591 men aged 59 to 70 years
still living in Hertfordshire.
28
Studies in Humans and CVD …
•Table II shows that concentrations are inversely related to
weight at one year of age.
Table II: Mean plasma fibrinogen in mm aged 59 to 70 years
Weight at one year (pounds)
Number of men
Fibrinogen (g/litre )*
≤ 18
37
3.21
-20
91
3.08
-22
177
3.14
-24
173
2.98
- 26
80
2.95
≥27
33
2.93
All
591
3.05
*Geometric mean values adjusted for age and cigarette smoking
29
Source: Barker & Martyn 1991
Studies in Humans and CVD …
• The inverse relation between systolic blood pressure and
birth weight present in the Hertfordshire men is shown in
table IV.
Table IV: Mean systolic blood pressure in men aged 59 to 70 years
Birthweight ( pounds)
Number of men
Systo!ic pressure
-55
31
169
- 65
95
166
-75
251
165
-85
233
163
-95
125
162
>95
56
162
All
791
164
30
Source: Barker & Martyn 1991
Studies in Humans and CVD …
• A similar relation has also been found in a national
sample of men and women at the age of 36 years.
• In contrast to plasma concentrations of fibrinogen and
rates of glucose intolerance, blood pressure in these
men is not related to weight at one year independently
of birth weight, nor is it related to adult height.
• This may indicate that the critical period when blood
pressure is sensitive to programming is during fetal
life rather than infancy.
31
Studies in Humans and CVD …
• These discoveries have implications both for the
pathogenesis of cardiovascular and other
diseases, and also for maternal and infant health
at specific locations.
• The relations between early growth and risk
factors and rates of disease are continuous.
32
Source: Barker & Martyn 1991
Studies in Humans and CVD …
• Plasma levels of fibrinogen (table II), the
prevalence of impaired glucose tolerance (table
III), and levels of systolic blood pressure (table
IV) fall progressively up to the highest values of
weight at one year or birthweight.
• If the criterion for successful fetal and infant
growth is adult health and longevity, we may no
longer be entitled to assume that a baby of average
birthweight and weight in infancy has necessarily
achieved its optimum weight.
33
Source: Barker & Martyn 1991
Geography of CVD …
• The results of these studies show that retarded growth
in fetal life and infancy is strongly related both to
mortality from cardiovascular disease and to adult
levels of some of its known risk factors.
• Any argument concerns the extent to which this
relation should be interpreted as being causal.
• In broad terms there are three possible explanations
for our findings.
• The first is that birthweight is merely a marker for
adverse environmental influences that act in later life.
34
Source: Barker & Martyn 1991
Geography of CVD …
• Although this interpretation can just be sustained
if one is prepared to view the ecological data in
isolation, it cannot account for the results of
follow up studies of individuals.
• In Hertfordshire birth weight was not associated
with social class, either at birth or currently.
35
Source: Barker & Martyn 1991
The Geographic Variation of Cancer
Incidence in Ontario
• Walter et al. 1994
• Examination of the spatial pattern of cancer
incidence in Ontario
• Cancer incidence were calculated for 22 cancer
sites in 49 counties in Ontario during 1976-1986.
36
The Geographic Variation of Cancer
Incidence in Ontario …
• The analysis reveal a number of cancers with
significant patterning of risk
• Further work is needed to relate the cancer data to
other information on potential life-style and
environmental factors
37
Geography of Cancer
• Stomach cancer
•
Strong spatial aggregation
was apparent in both sexes.
• Northern Ontario and some
areas near Lake Ontario had
high rates, while central
Ontario had lower rates
• Correlations between sexes
and over time were high
38
Geography of Cancer…
• Lung cancer.
• There were strong geographic
patterns for both sexes, with
high correlations between
sexes and over time.
• High rates were found in
large areas of the northeast
and east and in some counties
near Lake Ontario and in the
southwest
39
Geography of Cancer …
• Malignant melanoma of
the skin
• Female rates showed a
significant clustering
• Elevated risk in a group of
counties north of Toronto
40
Geography of Cancer
• Cervical Cancer
• Weak and dispersed
geographic pattern with
strong persistence over
time
• High rates tended to occur
in northern Ontario and
around Georgian Bay
• Rates in the southwest
were generally low
41
Geography of Cancer …
• Corpus uteri cancer
• Significant spatial pattern
with persistence over time
• Rates were low in northern
Ontario and generally high in
the southwest
42
Geography of Cancer …
• Prostatic cancer
• Strong spatial aggregation
• A risk gradient by latitude,
with low rates in the north and
high rates in the southwest
• Low correlation over time,
suggesting a transient or
recent effect
43
Geography of Cancer …
Major Known Risk Factors for Cancer Sites/Types with Evidence of Spatial Aggregation
Site/Type
Risk Factor(s)
Lip
Sunlight, tobacco
Esophagus
Tobacco, alcohol, diet low in fresh fruits and vegetables, low; socioeconomic
status (SES)
Stomach
Diet low in fresh fruits and vegetables, high intake of nitrates/ nitrites (e,g., in
water and preserved foods), low SES
Colon/Rectum
Diet high in fat, diet low in the fiber, high SES (colon)
Pancreas
Tobacco
Lung
Tobacco, occupational exposures (radon, asbestos, etc.), low SES
Bone
Ionizing radiation
Malignant melanoma
Ultraviolet radiation, susceptible phenotype (fair hair/eyes/ skin), high SES
Cervix uteri
Tobacco, sexual promiscuity (early age at coitus, multiple sex partners),
viruses (human papilloma virus, herpes simplex virus type 2), low SES
Corpus uteri
High levels of unopposed endogenous estrogens, use of exogenous estrogens,
use of combination of oral contraceptives, high SES
Prostate
Diet (possibly high in fat), high levels of androgens
Kidney
Tobacco, obesity
44
Geography of Cancer …
• Several cancers show regional variations, but the
explanation for the variability requires an
understanding of the possible impact of
environmental factors
• However, migration and latency are major
problems.
45