Estimating the burden of disease attributable to high cholesterol in South Africa in 2000.

Abstract

To estimate the burden of disease attributable to high cholesterol in adults aged 30 years and older in South Africa in 2000.
World Health Organization comparative risk assessment (CRA) methodology was followed. Small community studies were used to derive the prevalence by population group. Population-attributable fractions were calculated and applied to revised burden of disease estimates for the relevant disease categories for each population group. The total attributable burden for South Africa in 2000 was obtained by adding the burden attributed to high cholesterol for the four population groups. Monte Carlo simulation-modelling techniques were used for uncertainty analysis.
South Africa.
Black African, coloured, white and Indian adults aged 30 years and older.
Mortality and disability-adjusted life years (DALYs) from ischaemic heart disease (IHD) and ischaemic stroke.
Overall, about 59% of IHD and 29% of ischaemic stroke burden in adult males and females (30+ years) were attributable to high cholesterol (>or= 3.8 mmol/l), with marked variation by population group. High cholesterol was estimated to have caused 24,144 deaths (95% uncertainty interval 22,404 - 25,286) or 4.6% (95% uncertainty interval 4.3 - 4.9%) of all deaths in South Africa in 2000. Since most cholesterol-related cardiovascular disease events occurred in middle or old age, the loss of life years comprised a smaller proportion of the total: 222,923 DALYs (95% uncertainty interval 206,712 - 233,460) or 1.4% of all DALYs (95% uncertainty interval 1.3 - 1.4%) in South Africa in 2000.
High cholesterol is an important cardiovascular risk factor in all population groups in South Africa.

Full text

Page 1

August 2007, Vol. 97, No. 8 SAMJ
ORIGINAL ARTICLES
708 708
The value of abnormal blood lipids and apo-lipoprotein levels
to predict ischaemic heart disease (IHD) has been studied
for decades, with the initial focus shifting from studying the
relationship with total blood cholesterol (TC) to that of the
individual components associated with TC. TC is a composite
measure of the cholesterol content of lipoprotein particles such
as low-density lipoprotein cholesterol (LDLC) and high-density
lipoprotein cholesterol (HDLC). The dyslipidaemic patterns
of high levels of LDLC and/or low HDLC have been found
to impart high risk for developing IHD, and justify aggressive
treatment to reduce the risk.1 Recent data suggest that levels of
the actual apo-lipoproteins such as ApoB and ApoA1 are the
most powerful lipid predictors for IHD.2
In most developing countries the only lipid-related
population-based data available are TC levels. However,
TC data can be used effectively as a proxy measure of other
dyslipidaemic markers of risks for IHD, ischaemic stroke
and other vascular diseases. The risks of TC are continuous
and occur even in populations with TC levels much lower
than those experienced in North American and European
populations.3,4 The level of TC usually increases with age in
populations with a westernised lifestyle. High TC levels are
associated with a diet high in fat (particularly saturated animal
fat), with cholesterol in the diet, and with lack of fibre.
Cardiovascular disease (CVD) risk factors, including high
TC, are already prevalent in the South African population,
and are set to increase with growing urbanisation of the black
African population.5-7 (Population group classification is used
in this article to demonstrate differences in risk-factor profile
and the subsequent burden. Data are based on self-reported
categories according to the population group categories
used by Statistics South Africa. Mentioning such differences
allows for a more accurate estimate of the overall burden
and may assist in higher effectiveness of future interventions.
The authors do not subscribe to this classification for any
other purpose.) In the black African population currently
the predominant CVD is cerebrovascular disease (stroke).8
However, as the TC levels increase, the CVD pattern will
shift towards higher rates of IHD. It is already consistently
reported that the caseload of IHD is on the increase in the black
Estimating the burden of disease attributable to high
cholesterol in South Africa in 2000
Rosana Norman, Debbie Bradshaw, Krisela Steyn, Thomas Gaziano and the South African Comparative Risk Assessment
Collaborating Group
Burden of Disease Research Unit, Medical Research Council of South Africa, Tyger-
berg, Cape Town
Rosana Norman, PhD
Debbie Bradshaw, DPhil (Oxon)
Chronic Diseases of Lifestyle Research Unit, Medical Research Council of South
Africa and Department of Medicine, University of Cape Town
Krisela Steyn, MD, MSc, NED
Divisions of Cardiovascular and Social Medicine and Health Inequalities, Brigham
and Women’s Hospital, Harvard Medical School, Boston, USA
Thomas Gaziano, MD, MSc
Corresponding author: R Norman (rosana.norman@mrc.ac.za)
Objectives. To estimate the burden of disease attributable to high
cholesterol in adults aged 30 years and older in South Africa in
2000.
Design. World Health Organization comparative risk assessment
(CRA) methodology was followed. Small community studies
were used to derive the prevalence by population group.
Population-attributable fractions were calculated and applied
to revised burden of disease estimates for the relevant disease
categories for each population group. The total attributable
burden for South Africa in 2000 was obtained by adding the
burden attributed to high cholesterol for the four population
groups. Monte Carlo simulation-modelling techniques were
used for uncertainty analysis.
Setting. South Africa.
Subjects. Black African, coloured, white and Indian adults aged
30 years and older.
Outcome measures. Mortality and disability-adjusted life years
(DALYs) from ischaemic heart disease (IHD) and ischaemic
stroke.
Results. Overall, about 59% of IHD and 29% of ischaemic stroke
burden in adult males and females (30+ years) were attributable
to high cholesterol (≥ 3.8 mmol/l), with marked variation by
population group. High cholesterol was estimated to have
caused 24 144 deaths (95% uncertainty interval 22 404 - 25 286) or
4.6% (95% uncertainty interval 4.3 - 4.9%) of all deaths in South
Africa in 2000. Since most cholesterol-related cardiovascular
disease events occurred in middle or old age, the loss of life
years comprised a smaller proportion of the total: 222 923
DALYs (95% uncertainty interval 206 712 - 233 460) or 1.4% of
all DALYs (95% uncertainty interval 1.3 - 1.4%) in South Africa
in 2000.
Conclusions. High cholesterol is an important cardiovascular
risk factor in all population groups in South Africa.
S Afr Med J 2007; 97: 708-715.
high cholestrol.indd 7087/31/07 5:04:20 PM

Page 2

ORIGINAL ARTICLES
709
August 2007, Vol. 97, No. 8 SAMJ
African population, particularly in urban hospitals. It has been
suggested that the incidence of IHD in the urban black African
population may even be as high as in the white population.9
Nine community studies10-19 have been conducted in South
Africa, but there are no nationally representative data on TC
levels. This article aimed to use the community study data
to derive estimates of national prevalence of exposure to
high TC in adults aged 30 years and older by sex, age and
population group. It also aimed to quantify the adverse health
consequences associated with high TC by population group,
and to estimate the burden of disease attributed to high TC
by sex, population and age group in South Africa for the year
2000.
Methods
Using World Health Organization (WHO) comparative
risk assessment (CRA) methodology3,20 the disease burden
attributable to a particular risk factor is estimated by
comparing current local health status with a theoretical
minimum counterfactual with the lowest possible risk. The
attributable fraction of disease burden in a population is
determined by prevalence of exposure to the risk factor in the
population and the relative risk (RR) of disease occurrence
given exposure. Cholesterol was defined as a continuous
variable with mean and standard deviation (SD) values of total
serum cholesterol expressed in millimoles per litre of serum
(mmol/l).
The outcomes assessed were IHD and ischaemic stroke. IHD
was chosen in the WHO CRA study4 on the basis of clear and
consistent positive associations observed in cohort studies,
as well as on the evidence of reversibility in clinical trials
of treatments to lower cholesterol levels. There is a positive
association between cholesterol and ischaemic stroke, but
there is a qualitatively different association with haemorrhagic
stroke.4 As the endpoints had to be mapped to the South
African burden of disease study classification system, total
stroke (which includes both stroke subtypes) was used in this
analysis. It was necessary, however, to adjust stroke burden
by the proportions of total fatal and non-fatal stroke that were
ischaemic stroke, using the proportions by stroke subtype for
the WHO African subregion AFR-E, which includes South
Africa.4 Although cholesterol has been shown to be inversely
associated with cancer and chronic respiratory disease,
evidence suggests that these associations were rather due to
the effects of the diseases on cholesterol levels, and hence these
outcomes were not included in the analysis.4
Estimates of mean TC levels and SDs by age and sex for
adults aged 30 years and older were obtained by pooling
data from the nine available studies10-19 of randomly selected
subjects conducted between 1980 and 2000. The means and SDs
were weighted according to the number of observations of each
study and the SD also took into account the difference between
the means of the studies. The pooled TC mean was calculated
as m=Σωimi and the pooled SD as
(
1
=
∑
=
i
where ωi was a proportional weight based on the number of
subjects in each study and mi and sdi were the study mean and
SD respectively.
Owing to differences in prevalence of CVD in the four
population groups, the analysis was carried out separately for
each. Age- and sex-related population group-specific mean
cholesterol levels from these community studies were weighted
according to the size of the population in each age-sex group in
South Africa in 2000.21
Cholesterol values for those aged 70 - 79 years and > 80 years
were assumed to be the same as for the 60 - 69-year age group,
owing to the extremely limited availability of cholesterol
levels in community studies for adults older than 65. This may
have led to a slight overestimate in the white population, as
the association between age and cholesterol is non-linear.4 In
high-cholesterol regions, cholesterol levels for females increase
between the ages of 30 and 65 years, and then fall slightly. In
males, levels increase between the ages of 30 and 50 years, and
then flatten before declining slightly in older age.4 Lawes et al.4
did not assume a decline in older ages in the low-cholesterol
AFR-E region.
A comprehensive review of the literature by Lawes et
al.4 identified the Asia-Pacific Cohort Studies Collaboration
analysis of 1999 (APCSC 1999)22 as the primary data source
for RR estimates of IHD and ischaemic stroke. APCSC 1999
analyses offered greater sample size, and while other available
meta-analyses used tabular data, APCSC analyses were based
on individual participant data and provided age-specific
analyses of total stroke and stroke subtypes. These analyses
could therefore more reliably adjust for confounding and
provide more reliable estimates of hazard ratios.4
In this analysis we used risk ratios based on more recent
analyses of the APCSC data, which reflect a smoother
estimate of the attenuation of RRs across age, as shown in
Table I (S Vander Hoorn, University of Auckland – personal
communication, 2005). These same risk ratios were used across
population groups. Related health outcomes and ICD-1023
codes are also presented in Table I.
Similar to the case of blood pressure, regression dilution
bias must be taken into account when observational data
are used to estimate the association between cholesterol and
cardiovascular endpoints. This bias occurs as baseline or
‘one-off’ measures of TC are subject to random fluctuations,
due partly to the measurement process and partly to real but
temporary deviations at a single time-point from the ‘usual’
exposure level. The measured values have a wider distribution
than the ‘usual’ values, and with repeated measures result in
1) 1
−
) 1
−
(
) 1
−
0
+
∑
=
k
1
=
PAF
ii
−
k
i
ii
R R
+
0
=
i
p
i
R Rp
PAF
1p) 1(
) 1
−
=
(
0
=
√
∑
=
i
∑
i
k
i
k
ii
R Rp
R Rp
ωisdi
PAF
2 +
ωi(mi –m)2
1(
) 1
−
(
1
+
∑
∑
=
k
ii
k
i
ii
R R
R Rp
high cholestrol.indd 7097/31/07 5:04:22 PM

Page 3

August 2007, Vol. 97, No. 8 SAMJ
ORIGINAL ARTICLES
710
‘regression to the mean’ of values, whereby an initially extreme
observation tends to become less abnormal with replication.
Since this bias is accounted for in the estimates of RR, it also
needs to be reflected in a narrower distribution around the
estimated mean TC levels.24,25 A correction for regression
dilution resulting from the ‘time-dependent’ variation from
the usual level is made by multiplying the SDs across all age-
sex categories by a factor of 0.625 (S Vander Hoorn – personal
communication).
Customised MS Excel spreadsheets based on templates
used in the Clinical Trial Research Unit at the University of
Auckland (S Vander Hoorn – personal communication, 2005)
as well as Australian studies (T Vos, University of Queensland
– personal communication, 2005) were used to calculate the
attributable burden using a discrete version of the general
potential impact fraction (see below), taking into account
continuous risk factor disease exposures compared with
a theoretical minimum distribution (conferring the lowest
possible risk), on a categorical scale.
∑
=
k
R Rp
(
0
=
i
i=1
1
=
∑
=
i
where n = the number of exposure categories; Pi = the
proportion of the population in exposure category i; RRi =
the RR for exposure category i; and P’i = the proportion of
the population in exposure category i in the counterfactual
distribution.
The theoretical minimum exposure distribution is zero in
most cases, since zero exposure reflects minimum risk (e.g.
no smoking). For TC, however, zero exposure would be
physiologically impossible, and therefore is an inappropriate
choice as the theoretical minimum. Lawes et al.4 examined data
from ‘unacculturated’ populations which included hunter-
gatherer societies where mean values were as low as 3.0 - 3.5
mmol/l. With diets low in salt and animal fat, the prevalence
of CVD is very low in these ‘low-cholesterol’ populations. In
populations such as China, data suggest that the relationship
between cholesterol and cardiovascular endpoints may
continue, without a threshold, below cholesterol levels of about
4.0 mmol/l. Based on this evidence, the theoretical minimum
of usual cholesterol was set at a mean of 3.8 mmol/l (SD 0.6
mmol/l – ‘usual’), for all age, sex and population groups.4
PAFs were then applied to revised South African burden of
disease estimates for 2000 for each population group,21 deaths,
premature mortality or years of life lost (YLL), years lived with
disability (YLD), and disability-adjusted life years (DALYs) for
the relevant disease categories, to calculate attributable burden.
The total attributable burden for South Africa in 2000 was
obtained by adding the burden attributed to high cholesterol
for the four population groups.
Monte Carlo simulation-modelling techniques were used
to present uncertainty ranges around point estimates that
reflect all the main sources of uncertainty in the calculations.
We used the @RISK software version 4.5 for Excel,26 which
allows multiple recalculations of a spreadsheet, each time
choosing a value from distributions defined for input variables.
Normal probability distributions were specified around the
mean TC levels by age, sex and population group. For the RR
input variables a normal distribution was specified, with the
natural logarithm of the RR estimates as the entered means
of the distribution and the standard errors of these estimates.
For each of the output variables (namely attributable burden
as a percentage of total burden in South Africa, 2000), 95%
uncertainty intervals were calculated bounded by the 2.5th and
97.5th percentiles of the 2000 iteration values generated.
Results
Table II shows the estimated mean TC levels, which were
highest in the white population group, followed closely by
the coloured and Indian groups. The black African population
had the lowest mean levels for males and females at all ages,
with levels in urban areas only slightly (but not significantly)
higher than rural levels. Mean TC levels in women were higher
than in men above 45 years of age in all population groups. In
the youngest age group, however, levels in men were higher
than in women in the Indian, coloured and white population
groups. TC levels in white women increased between the
ages of 30 and 65 years, and were then assumed to remain flat
– although they should fall slightly afterwards. In white males,
levels increased between the ages of 30 and 50 years and then
flattened, although they are also supposed to decline slightly
in older age groups. Similar age patterns were observed in
Indian and coloured males and females, but not in the black
African population. In the black African population the
increases between the 30 - 44-year and 45 - 59-year age groups
were not as marked as in the other population groups, and in
black African men the levels appeared to continue increasing
between ages of 30 and 65 years.
1
−
) 1
−
1
+
i
R R
0
=
(
) 1
−
(
i
p
P'i RRi
(
0
1
+
=
∑
PAF
∑
=
i
ii
i
k
p
i
ii
=
R R
−
i
p
R R
PAF
PAF
PAF
n
k
i=1Pi RRi –
1
=
i
∑
) 1
−
Pi RRi
(
i
p
∑
i
) 1
i=1
=
k
(
=
∑
k
ii
i
R Rp
n
k
1) 1
−
(
) 1
−
( R R
0
+
=
k
ii
R Rp
n
k
1) 1
) 1
−
1
+
∑
i
ii
i
p
R R
PAF
Table I. Age-specific hazard ratios (95% CI) for 1 mmol/l difference in usual total cholesterol
Related health outcome (ICD-10)23
Ischaemic heart disease (I20-I25) 0.27 (0.20 - 0.38)
Ischaemic stroke (I63) 0.61 (0.32 - 1.18)

Source: S Vander Hoorn, University of Auckland – personal communication, 2005; more recent reanalysis of Asia-Pacific Cohort Studies Collaboration (APCSC)22 data.
95% CI = 95% confidence interval.
30 - 44 yrs 45 - 59 yrs
0.48 (0.44 - 0.52)
0.69 (0.59 - 0.80)
60 - 69 yrs
0.64 (0.62 - 0.67)
0.76 (0.64 - 0.90)
70 - 79 yrs
0.70 (0.66 - 0.75)
0.82 (0.68 - 0.99)
80+ yrs
0.70 (0.67 - 0.73)
0.90 (0.75 - 1.08)
high cholestrol.indd 7107/31/07 5:04:23 PM

Page 4

ORIGINAL ARTICLES
711
August 2007, Vol. 97, No. 8 SAMJ
Table II. Estimates of age-specific mean total cholesterol levels and standard deviations (SDs) (mmol/l) for males and females
by population group, South Africa, 2000
Age groups (yrs)
Population Sex Parameter 30 - 44 45 - 59 60 - 69
Urban black African* Males Mean 4.4 4.5 4.7
SD† 0.8 0.7 0.7
Females Mean 4.5 5.0 5.1
SD† 0.6 0.7 0.6
Rural black African* Males Mean 4.4 4.6 4.6
SD† 0.7 0.6 0.7
Females Mean 4.4 4.7 5.0
SD† 0.6 0.7 0.7
Black African Males Mean 4.4 4.5 4.7
SD† 0.7 0.7 0.7
Females Mean 4.4 4.9 5.1
SD† 0.6 0.7 0.7
Coloured Males Mean 5.7 5.8 5.6
SD† 0.8 0.8 0.7
Females Mean 5.3 6.2 6.4
SD† 0.7 0.8 0.7
White Males Mean 5.8 6.3 6.2
SD† 0.9 0.8 0.7
Females Mean 5.5 6.4 7.2
SD† 0.7 0.8 0.9

Asian/Indian Males Mean 5.8 6.1 5.6
SD† 0.8 0.8 0.7
Females Mean 5.3 5.9 6.1
SD† 0.7 0.8 0.8
South Africa* Males Mean 4.8 5.0 5.2
SD† 0.8 0.9 0.8
Females Mean 4.6 5.3 5.7
SD† 0.7 0.9 0.9
Data sources
Study site Year
Cape Town (Western Cape)12 1990
Mangaung (Free State)10‡ 1990/1 758
Durban (KwaZulu-Natal)13 1986
70 - 79 80+
4.7
0.7
5.1
0.6
4.6
0.7
5.0
0.7
4.7
0.7
5.1
0.7
5.6
0.7
6.4
0.7
6.2
0.7
7.2
0.9

5.6
0.7
6.1
0.8
5.2
0.8
5.7
0.9
N
665
4.7
0.7
5.1
0.6
4.6
0.7
5.0
0.7
4.7
0.7
5.1
0.7
5.6
0.7
6.4
0.7
6.2
0.7
7.2
0.9

5.6
0.7
6.1
0.8
5.2
0.8
5.7
0.9
232
QwaQwa (Free State)10‡ 1990/1 853
Dikgale (Limpopo)11§ 1997/8 1 494
Weighted estimate
Cape Town (Western Cape)14 1982
Mamre (Western Cape)15 1996
779
646
Durban (KwaZulu-Natal)17‡ 1987
Robertson, Swellendam, 1983 2 722
Riversdal (Western Cape)18¶
Riversdal, Caledon, Bredasdorp 1993
(Western Cape)19II
Durban (KwaZulu-Natal)16‡ 1984/6
354
327
601
Weighted estimate
*Rural black African, urban black African and South African average estimates are presented but were not used in the calculations.
†Corrected for regression dilution resulting from time-dependent change in ‘usual’ level (ratio = 0.625).
‡Oldest age category in study is 55 - 69 years.
§Data of Dikgale study extracted from data from reference 11, youngest age group, 30 - 34 years.
¶Data for men of the CORIS study extracted from data from reference 18.
IIAge group 35 - 44 years.
Table III. Prevalence with serum cholesterol values above 5 mmol/l* by population group, age and sex, South Africa, 2000
Proportion (%) with serum cholesterol ≥ 5 mmol/l
Population group Age group (yrs) Males
Black African 30 - 44 21.5
45 - 59 24.5
60+ 32.2
30+ 23.8
Coloured 30 - 44 80.8
45 - 59 84.5
60+ 79.6
30+ 81.7
White 30 - 44 83.9
45 - 59 93.8
60+ 96.0
30+ 90.0
Asian/Indian 30 - 44 84.9
45 - 59 92.3
60+ 78.4
30+ 86.5
South Africa 30 - 44 38.9
45 - 59 50.5
60+ 58.0
30+ 45.2
Females
17.4
43.5
54.1
30.9
66.9
94.8
97.9
79.9
75.0
95.3
99.4
88.4
65.1
87.8
90.3
77.4
30.3
63.2
78.7
49.6
Persons
19.4
34.5
44.9
27.6
73.6
90.0
90.4
80.7
79.5
94.6
97.9
89.2
74.9
89.9
85.2
81.7
34.4
57.4
70.7
47.6
*Current clinical cut-off for hypercholesterolaemia.
high cholestrol.indd 7117/31/07 5:04:23 PM

Page 5

August 2007, Vol. 97, No. 8 SAMJ
ORIGINAL ARTICLES
712
The estimated prevalence of hypercholesterolaemia (TC ≥
5 mmol/l, the current clinical cut-off) by population group
in the age groups 30 years and older is shown in Table III.
The differences between the prevalence rates in the black
African population and the other three groups are striking,
and are consistent with the lower levels of urbanisation and
concomitant incomplete adoption of Western lifestyles in this
group.
The PAFs for high cholesterol for males, females and adults
of 30+ years are shown in Table IV. Attributable fractions were
higher in females for all related outcomes and in all population
groups. Overall, about 59% of IHD and 29% of ischaemic
stroke burden in adult males and females (30+ years of age)
were attributable to high cholesterol (≥ 3.8 mmol/l).
High cholesterol was estimated to cause 24 144 deaths (95%
uncertainty interval 22 404 - 25 286) or 4.6% (95% uncertainty
interval 4.3 - 4.9%) of all deaths in South Africa in 2000.
There are more female than male attributable deaths, with
attributable deaths from ischaemic stroke in females double
those in males (Table IV). Since most cholesterol-related CVD
events occurred in middle or old age, the loss of life years
comprises a smaller proportion of the total: 222 923 DALYs
(95% uncertainty interval 206 712 - 233 460) or 1.4% of all
DALYs (95% uncertainty interval 1.3 - 1.4%) in South Africa in
2000.
Age-standardised attributable mortality rates by population
group are presented in Fig. 1. Large population group
differences were observed, with the highest rates seen in
Indian males and females, followed by white males and
females. Coloured adults had intermediate rates, with very
low rates observed in the black African population group. It
is interesting to note that in the Indian and white groups, age-
standardised cholesterol-attributable mortality rates in males
were higher than in females. However, in the black African and
coloured groups the rates in females were slightly higher than
in males.
The contribution of IHD and ischaemic stroke to total
attributable burden by population group is shown in Fig.
2. While ischaemic stroke and IHD accounted for similar
proportions of total burden in the black African population
group, IHD accounted for the majority of the burden in
the Indian and white population groups, with a very small
contribution from ischaemic stroke.
Discussion
Overall, the results suggest that a considerable proportion
of CVD in South Africa can be attributed to non-optimal
cholesterol levels, which translates into a high proportion of
deaths and disability. Worldwide, 56% of IHD mortality and
disease burden was attributable to cholesterol levels of more
than 3.8 mmol/l, which translated to 3.6 million deaths in the
year 2000.4 In South Africa, 59% of IHD was attributable to
raised cholesterol. In addition, 29% of ischaemic stroke was
attributable to raised cholesterol, which compares with the
worldwide estimate of 32% (range 25 - 45% by subregion).4
Overall, 4.4 million deaths (about 7.9% of the total) and 40.4
million DALYs (2.8% of the total) worldwide were estimated
to be due to non-optimal cholesterol levels. About 40% of the
cholesterol-related attributable burden occurred in developed
subregions, 20% in low-mortality developing sub-regions, and
a further 40% in high-mortality developing subregions. The
relative impact of attributable deaths and DALYs was highest
in the European subregions, with the highest mean cholesterol
levels.4
40
135
220
299
47
144
152
232
0
50
100
150
200
250
300
350
Black AfricanColouredWhite
Female
Asian/Indian
Age standardised mortality rate per 100 000
Male
Fig. 1. Age-standardised cholesterol-attributable mortality rates by
population group and sex, South Africa, 2000.
Attributable DALYs= 21 990
Asian/Indian persons
Ischaemic
stroke
8.6%
Ischaemic
heart disease
91.4%
Attributable DALYs= 81 556
White persons
stroke
10.2%
stroke
10.2%
10.2%
Ischaemic heart
disease
89.8%
South Africa, 2000.
Ischaemic
stroke
10.2%
Attributable DALYs= 29 652
Coloured persons
Ischaemic
stroke
26.0%
26.0%
26.0%
Ischaemic
heart disease
74.0%
Ischaemic
stroke
26.0%
Attributable DALYs= 89 726
Black African persons
Ischaemic heart
disease
58.1%
Ischaemic
stroke
41.9%
Attributable DALYs= 21 990
Asian/Indian persons
Ischaemic
stroke
8.6%
Ischaemic
heart disease
91.4%
Attributable DALYs= 81 556
White persons
Ischaemic
stroke
10.2%
Ischaemic heart
disease
89.8%
Ischaemic
stroke
10.2%
Attributable DALYs= 29 652
Coloured persons
Ischaemic
heart diseas
74.0%
Ischaemic
stroke
26.0%
Attributable DALYs= 89 726
Black African persons
Ischaemic heart
disease
58.1%
Ischaemic
stroke
41.9%
Ischaemic heart
disease
58.1%
Attributable DALYs= 21 990
Asian/Indian persons
Ischaemic
stroke
8.6%
Ischaemic
heart disease
91.4%
Attributable DALYs= 81 556
White persons
Ischaemic heart
disease
89.8%
Attributable DALYs= 29 652
Coloured persons
Ischaemic
heart disease
74.0%
Ischaemic
stroke
26.0%
Attributable DALYs= 89 726
Black African persons
Ischaemic heart
disease
58.1%
Ischaemic
stroke
41.9%
Ischaemic heart
disease
58.1%
58.1%
Attributable DALYs= 21 990
Asian/Indian persons
Ischaemic
stroke
8.6%
stroke
8.6%
8.6%
Ischaemic
heart disease
91.4%
heart disease
91.4%
91.4%
Attributable DALYs= 81 556
White persons
Attributable DALYs= 81 556
White persons
White persons
Ischaemic heart
disease
89.8%
disease
89.8%
89.8%
Ischaemic
Ischaemic
Ischaemic
stroke
Attributable DALYs= 29 652
Coloured persons
Coloured persons
Coloured persons
Ischaemic
heart disease
74.0%
74.0%
74.0%
Attributable DALYs= 89 726
Black African persons
Black African persons
Black African persons
Ischaemic
stroke
41.9%
41.9%
41.9%
Fig. 2. Burden attributable to high cholesterol by population group,
Attributable DALYs= 21 990
Asian/Indian persons
Ischaemic
Ischaemic
stroke
Ischaemic
Ischaemic
heart disease
Ischaemic heart
Ischaemic heart
disease
Attributable DALYs= 29 652
Attributable DALYs= 29 652
Ischaemic
heart disease
heart disease
Ischaemic
stroke
stroke
Attributable DALYs= 89 726
Attributable DALYs= 89 726
Ischaemic heart
disease
Ischaemic
stroke
stroke
Attributable DALYs= 21 990
Asian/Indian persons
Attributable DALYs= 81 556
Ischaemic
Ischaemic
Ischaemic
high cholestrol.indd 7127/31/07 5:04:24 PM