Journal of Clinical Lipidology (2018) -, -–-

Original Article

Prevalence, patterns, and associations

of dyslipidemia among Sri Lankan
adults—Sri Lanka Diabetes and
Cardiovascular Study in 2005–2006
Prasad Katulanda, MD, DPhil (Oxon), Harsha Anuruddhika Dissanayake, MBBS*,
S. D. Neomal De Silva, MBBS, Gaya Wijeweera Katulanda, MD,
Isurujith Kongala Liyanage, MD, Godwin Roger Constantine, MD, Rezvi Sheriff, MD,
David R. Matthews, MD, PhD

Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, London, UK (Drs Katulanda and
Matthews); Diabetes Research Unit, Department of Clinical Medicine, University of Colombo, Colombo, Sri Lanka (Drs
Katulanda, Dissanayake, De Silva, Liyanage, Constantine, and Sheriff); and Medical Research Institute, Colombo, Sri
Lanka (Dr Katulanda)

KEYWORDS: BACKGROUND: Dyslipidemia is a major risk factor for cardiovascular disease. Prevalence patterns
Dyslipidemia; and determinants of dyslipidemia in Sri Lanka are unkown.
Sri Lanka; OBJECTIVES: We aimed to determine the prevalence and correlates of dyslipidemia among Sri Lan-
Diabetes; kan adults.
Cardiovascular risk; METHODS: A nationally representative sample was recruited by multistage random cluster sampling
Metabolic syndrome; in Sri Lanka Diabetes and Cardiovascular Study, a cross-sectional study. Data collected by an
Obesity interviewer-administered questionnaire, physical examination, anthropometric measurements lipid
analysis from take 12-hour fasting blood samples were used.
RESULTS: Among 4451 participants 60.5% were women and mean age was 46 years. Mean (stan-
dard deviation) total cholesterol (TC), high-density lipoprotein cholesterol (HDLC), low-density lipo-
protein cholesterol (LDLC), triglycerides (TGs), and TC/HDLC levels were 206.7 mg/dL (643.5),
46.8 mg/dL (610.6), 135.5 mg/dL (637.6), 121.7 mg/dL (666.8), and 4.6 (61.1), respectively.
Women had higher mean TC, HDLC, LDLC, and TG values compared to men across all age groups.
Mean TC, LDLC, and TGs increased with age in both genders; 77.4% of participants had some form of
dyslipidemia. Low HDLC was the commonest type (49.6%) of dyslipidemia. Increasing age, female
sex, living in urban sector, high body mass index, central obesity, diabetes, hypertension, insufficient
physical activity, and smoking were associated with having some form of dyslipidemia.
CONCLUSION: Three in four Sri Lankan adults have some form of dyslipidemia. Physical inactivity,
obesity, hypertension, and diabetes are the leading modifiable risk factors.
Ó 2018 National Lipid Association. All rights reserved.

Conflicts of interest: The authors declare no conflict of interest. E-mail address: dissanayakeha@gmail.com
* Corresponding author. Diabetes Research Unit, Department of Clin- Submitted August 30, 2017. Accepted for publication January 10,
ical Medicine, Faculty of Medicine, University of Colombo, Colombo 08, 2018.
Sri Lanka.

1933-2874/Ó 2018 National Lipid Association. All rights reserved.

https://doi.org/10.1016/j.jacl.2018.01.006
2 Journal of Clinical Lipidology, Vol -, No -, - 2018
Background
Cardiovascular disease (CVD) remains the leading cause
of hospital mortality in Sri Lanka.1 Dyslipidemia is a major
risk factor for development of CVD.2 Treatment of lipid ab-
normalities is of proven benefit in reducing the morbidity
and mortality associated with CVD.3,4 Several studies
have identified low-density lipoprotein cholesterol
(LDLC) as the main atherogenic lipoprotein, which has
the strongest association with development of CVD.5
National surveys on dyslipidemia in Australia and
United States have shown prevalence ranging from 48%
to 53%.6,7 A USA national survey showed an association of
dyslipidemia with obesity.7 According to a British study,
dyslipidemia was significantly associated with diet, gender,
and age.8 A study from Nigeria and African lower middle-
income countries revealed that 23% of the population had
elevated total cholesterol (TC), 51% had elevated LDLC,
and 60% had low high-density lipoprotein cholesterol
(HDLC) with females recording better overall lipid profile9
when interpreted according to the National Cholesterol Ed-
ucation Programme/Adult Treatment Panel III (NCEP/ATP
III) criteria for definition and risk classification.
Data from Sri Lanka on dyslipidemia are limited. In
1994, a study in central province on prevalence of
cardiovascular risk factors found median serum TC to be
189.6 mg/dL and median HDLC to be 37.6 mg/dL.10
Objectives
Sri Lanka has no countrywide data on patterns and
distribution of lipids. Identifying the prevalence and nature
of dyslipidemia among Sri Lankans would help determine
risk of CVD. Therefore, we aimed to determine prevalence,
patterns, and determinants of dyslipidemia in a large
representative randomly selected adult population in Sri
Lanka.
Methods
Sri Lanka Diabetes and Cardiovascular Study (SLDCS)
was a cross-sectional study conducted by the Diabetes
Research Unit of the Faculty of Medicine, University of
Colombo, and the Oxford Centre for Diabetes Endocri-
nology and Metabolism of the University of Oxford, UK.
Ethical approval was obtained from the Ethics Review
Committee of the Faculty of Medicine, University of
Colombo. All participants provided informed written con-
sent. Data collection was conducted between August 2005
and September 2006. Detailed methods have been previ-
ously published.11
Study population
SLDCS was conducted in 7 of all 9 provinces in Sri
Lanka. Northern and the Eastern provinces (14% of the
total population) were excluded due to insurgency pre-
vailed. The total sample frame was approximately 14
million people living in 12,018 ‘‘village officer’’ units. A
multistage random cluster sampling technique was used to
select a nationally representative sample of 5000 noninsti-
tutionalized adults aged 18 years or above. Those who were
pregnant (n 5 26), acutely ill (febrile illness, immobility
following trauma, or recent hospitalization) (n 5 19) or
who declined participation (n 5 222), and those on any
form of lipid-lowering medication (n 5 207) were
excluded.
Data collection
Temporary data collection centers were established
within or in close proximity to each cluster. The selected
subjects were advised to visit the centers in the morning
between 7.30 AM and 9.00 AM after an overnight fast of 12–
14 hours. They were advised to follow an unrestricted diet
and to continue with usual physical activities for at least 3
days before visit to data collection centers. An interviewer-
administered, structured, and precoded questionnaire was
used for data collection. Data on demography, medical his-
tory, medication usage, physical activities, and alcohol and
smoking status were recorded. Anthropometric measure-
ments were noted, and fasting blood samples were drawn.
Serum was stored at 220 C. TC, HDLC, and triglycerides
(TGs) were measured by enzymatic photometric methods
using a Hitachi 704 chemical autoanalyzer (Roche Diag-
nostics, Mannheim, Germany) in the Reproductive and
Endocrinology Laboratory, Faculty of Medicine, University
of Colombo, Sri Lanka. Internal quality control samples
were run daily and participated in a monthly external qual-
ity assurance program. LDLC was calculated using the
Friedewald formula.
Statistical analysis
Central tendencies of continuous variables were ex-
pressed as mean 6 standard deviation, whereas data on
categorical variables were expressed as percentages or
frequencies. Determinants of dyslipidemias were analyzed
applying multivariate logistic regression. Comparison of
continuous variables in discrete groups was done using
Student’s t-test or analysis of variance. All the statistical
analyses were conducted with SPSS 16.0 statistical soft-
ware package (SPSS, Inc., Chicago, IL), and P
values , .05 were considered to be statistically significant.
Definitions
Lipid profiles were interpreted according to NCEP/ATP
III criteria (Tables 1 and 2). Patients with diagnosed dia-
betes (by a registered medical practitioner, confirmed by re-
view of medical records) or newly diagnosed (based on
American Diabetes Association and WHO criteria) were
considered as having diabetes. Level of physical activity
Katulanda et al Dyslipidemia among Sri Lankan adults 3

Table 1 Definition of cholesterol categories according to National Cholesterol Education Programme/Adult Treatment Panel III
guidelines
Lipid category Total cholesterol (mg/dL) LDL cholesterol (mg/dL) Triglycerides (mg/dL) HDL cholesterol (mg/dL)
Normal/desirable ,200 ,100 ,150 .40
Near optimal – 100–129 – –
Borderline high 200–239 130–159 150–199 –
High $240 160–189 200–499 –
Very high – $190 $500 –
HDL, high-density lipoprotein; LDL, low-density lipoprotein.

was estimated and classified according to International
Physical Activity Questionnaire. Urban, rural, and estate
sector definitions were in accordance with those of the
Sri Lankan government. Hypertension was diagnosed based
on previous diagnosis made by a registered medical practi-
tioner or based on newly detected elevated blood pressure
(systolic . 140 mmHg and/or diastolic . 90 mmHg on
at least 2 occasions) (detailed elsewhere11). Asian cutoffs
for body mass index (BMI: underweight , 18.5, normal
18.0–22.9, overweight 23.0–27.5, and obesity . 27.5 kg/
m2) and waist circumference (WC: .90 cm in men and
.80 cm in women considered high) were used in analysis.
Results
Of the 5000 invited subjects, 4532 participated in the
study (response rate 91%). This report is based on 4451
respondents, excluding 81 with incomplete data on lipid
parameters. Mean age of participants was 46.1 (615.1)
years, and 60.5% were women. Demographic and anthro-
pometric characteristics and lipid levels are summarized in
Table 3. Women had higher mean TC, LDLC, and non-
HDLC levels, whereas men had higher TG and lower
HDLC level.
Prevalence of dyslipidemia
A high prevalence of dyslipidemia of 77.4% was
identified in the studied population according to NCEP/
ATP III criteria. The prevalence was higher in women
(80.7% vs 73.5% in men) overall and across all age groups.
Figure 1 summarizes the prevalence of different types of
dyslipidemias. Low HDLC was the commonest type of dys-
lipidemia affecting 49.6% of the participants. Nearly one-
third of men and two-thirds of women had low HDLC,
and these proportions were consistent across age groups.
High LDLC and high TGs were prevalent in 46% and
23%, respectively. Mixed dyslipidemia (high TGs, high
LDLC, and low HDLC) was prevalent in 7.6% of the
participants.
Mean non-HDLC levels were high in both men and
women (Table 3). Forty-four percent of the participants had
non-HDLC more than 160 mg/dL, and in 19.5%, it was
more than 190 mg/dL.
Determinants of dyslipidemia
Prevalence of high LDLC increased with age (Fig. 2). It
was common in women aged 50 years and above, whereas
it was common in men at younger age. Prevalence of high
TG increased to reach a peak in the 50- to 59-year-old
group and declined after that. Hypertriglyceridemia is com-
mon among men across all age groups except those aged
70 years and above.
Although women had higher HDLC level across all age
groups, those HDLC values were below the cutoff of
optimum value (50 mg/dL). Among those aged more than
50 years, women had significantly higher TC and LDLC

Table 2 Definition of dyslipidemia based on National Cholesterol Education Programme/Adult Treatment Panel III guidelines
Cholesterol type Measurement ATP III category
LDL cholesterol
CHD or CHD risk equivalent or 10-y Framingham risk (20%) #100 mg/dL At goal
CHD or CHD risk equivalent or 10-y Framingham risk (20%) $100 mg/dL Above goal
2 risk factors or 10-y Framingham risk (20%) $130 mg/dL Above goal
0–1 risk factor $160 mg/dL Above goal
HDL cholesterol ,40 mg/dL (in men) Low
,50 mg/dL (in women)
Triglycerides ,150 mg/dL Normal
150–199 mg/dL Borderline high
200–400 mg/dL High
ATP III, Adult Treatment Panel III; CHD, coronary heart disease; HDL, high-density lipoprotein; LDL, low-density lipoprotein.
4 Journal of Clinical Lipidology, Vol -, No -, - 2018

Table 3 Demographic characteristics and mean (SD) BMI, waist circumference, and lipid levels of participants (N 5 4451)
P (comparing
men and
Men Women women) Total
Number (% of total population) 1758 (39.5%) 2693 (60.5%) – 4451
Age (y)* 46.3 (615.8; 18.0–89.0) 46.0 (614.6; 18.0–90.0) .547 46.1 (615.1)
Sector of living (%)
Estate (%) 92 (5.1) 107 (3.9) – 199 (4.4)
Rural (%) 1371 (77.3) 2126 (78.4) – 3497 (78.0)
Urban (%) 309 (17.5) 480 (17.7) – 789 (17.6)
Diabetes mellitus (%) 200 (11.3) 336 (12.4) .009 536 (12.0)
Hypertension (%) 477 (27.1) 742 (27.5) .512 1219 (27.4)
Current smoking (%) 684 (38.6) 2 (0.1) ,.001 686 (15.3)
Physical activity
Insufficient (%) 258 (14.6) 237 (8.7) – 495 (11.1)
Moderately active (%) 459 (25.9) 838 (30.9) – 1297 (28.9)
Highly active (%) 1055 (59.5) 1638 (60.4) – 2693 (60)
BMI (kg m22)* 21.1 (63.7; 14.1–39.3) 22.2 (64.5; 15.1–42.3) ,.001 21.8 (64.2)
Waist circumference (cm)* 78.1 (611.0; 52.0–121.0) 76.8 (612.2; 50.0–124.0) ,.001 77.3 (611.8)
Total cholesterol (mg/dL)* 202.1 (642.9; 98.0–453.0) 209.7 (629.1; 82.0–477.0) ,.001 206.7 (643.5)
LDLC (mg/dL)* 130.7 (636.6; 26.2–312.6) 138.5 (637.9; 29.0–355.0) ,.001 135.5 (637.6)
HDLC (mg/dL)* 44.6 (610.4; 21.0–107.0) 48.2 (610.6; 22.0–101.0) ,.001 46.8 (610.6)
Triglycerides (mg/dL)* 132.8 (673.7; 38.0–638.0) 114.4 (660.8; 31.0–1000.0) ,.001 121.7 (666.8)
Non-HDLC (mg/dL)* 158.1 (638.8; 55.0–396.0) 162.5 (625.2; 51.0–407.0) ,.001 160.3 (643.3)
TC/HDLC* 4.5 (61.3; 1.7–9.3) 4.35 (61.4; 2.0–16.0) .020 4.4 (61.4)
BMI, body mass index; HDLC, high-density lipoprotein cholesterol; LDLC, LDL, low-density lipoprotein cholesterol; SD, standard deviation.
*Values indicated are mean (6SD; range).

than males (P , .005). Among younger participants, men
had significantly higher mean TGs than females
(P , .005) and higher TC and LDLC levels; however, the
latter two were statistically insignificant. Table 4 summarizes
the mean lipid levels in men and women across age groups.
Adjusted multivariate logistic regression analysis
showed that high LDLC was associated with increasing
age, female sex, living in urban sector, increasing BMI and
WC, diabetes, hypertension, and current smoking, but not
with physical activity level (Table 5). Low HDLC was asso-
ciated with female sex, increasing BMI and WC, diabetes,
and urban living, but not with hypertension, smoking status,
or physical activity. Hypertriglyceridemia was associated
with male sex, increasing BMI, diabetes, hypertension, cur-
rent smoking, and insufficient level of physical activity, but
not with age. Advanced age, female sex, living in urban
sector, increasing BMI, and hypertension were risk factors
for high TC. No lipid abnormalities were associated with
education level, occupation, income category, or current
alcohol consumption in adjusted and nonadjusted models
(data not shown).

Discussion
LOW
HDL 49.6 In this article, we describe prevalence of dyslipidemia and
its correlates in a nationally representative population of Sri
Lankan adults. This is the first such national-level study
22.2
conducted in Sri Lanka and one of the first studies of this
extent in a large South Asian population. Demographic
5.7 14.1 characteristics of our study show few differences to the
general population statistics of Sri Lanka where mean age is
7.6
36.7 years, urban population is 13.2%, and female percent-
18.
age is 52.2%. Disparity in age and sex distribution is most
3.5
HIGH 6.2 HIGH likely due to exclusion of those aged below 18 years.
TG LDL

23 46
Prevalence
Figure 1 Percentage prevalence of isolated and mixed dyslipi-
demias by types. HDL, high-density lipoprotein; LDL, low- Remarkably high prevalence of dyslipidemia was
density lipoprotein; TG, triglyceride. observed among Sri Lankans according to this study, which
Katulanda et al Dyslipidemia among Sri Lankan adults 5

A 80

(0.9)
(1.2)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Total
60

4.0
4.2
4.5
4.6
4.7
4.7
4.6
% 40

(0.9)
(1.4)
(1.0)
(1.1)
(1.1)
(1.1)
(1.1)
Men
20 Women

3.8
4.0
4.3
4.5
4.7
4.8
4.7
F
0

(0.9)
(1.0)
(1.1)
(1.1)
(1.2)
(1.1)
(1.0)
TC:HDLC
< 20 20 - 29 30 - 39 40 - 49 50 - 59 60 - 69 > 69

4.2
4.4
4.7
4.8
4.8
4.6
4.5
Age (years)
B

F, females; HDLC, high-density lipoprotein cholesterol; LDLC, low-density lipoprotein cholesterol; M, males; TC, total cholesterol; TG, triglyceride; SD, standard deviation.
80

(36.5)
(48.7)
(60.6)
(73.4)
(75.0)
(60.7)
(60.2)
60

92.0
94.7
114.8
126.1
137.2
129.1
122.8
Total
% 40
Men

(25.0)
(38.4)
(47.9)
(65.6)
(66.9)
(56.7)
(69.7)
20 Women
0

83.4
85.0
100.8
115.0
131.1
130.3
129.9
< 20 20 - 29 30 - 39 40 - 49 50 - 59 60 - 69 > 69

F
Age (years)

(44.4)
(57.6)
(71.1)
(81.7)
(85.7)
(66.7)
(45.9)
C 40
35

101.0
108.2
137.1
144.6
147.1
127.3
114.7
30

TG
M
25

(13.1)
(10.5)
(10.0)
(10.6)
(11.3)
(10.8)
(10.1)
% 20
15 Men

Total
45.6
45.8
45.8
46.3
48.2
47.5
47.9
10 Women
5

(13.0)
(11.1)
(10.2)
(10.2)
(11.0)
(10.1)
(10.2)
0
< 20 20 - 29 30 - 39 40 - 49 50 - 59 60 - 69 > 69

48.5
48.0
46.9
47.3
50.0
48.6
49.2
Age (years)

F
Figure 2 Prevalence of (A) low HDL cholesterol, (B) high LDL

(12.9)

(9.52)
(10.6)
(11.3)
(11.6)
(8.7)

(9.8)
cholesterol, and (C) hypertriglyceridemia in men and women of
different age groups. Number of participants in each age category
HDLC

42.6
42.7
44.2
44.6
45.3
45.7
46.3
in ascending order: males: 22,225, 305, 383, 321, 191, 150; fe-
M

males: 23, 333, 487, 614, 512, 313, 185. HDL, high-density lipo-
(26.0)
(29.3)
(34.1)
(36.3)
(41.0)
(40.9)
(35.9)
protein; LDL, low-density lipoprotein.
108.9
118.7
130.4
136.0
144.0
145.0
140.6
Total
Lipid parameters of men and women based on age categories

exceeds the rates from most other regional and non-Asian

(28.2)
(29.1)
(32.9)
(35.0)
(40.7)
(40.6)
(37.3)

countries. For example, 77.4% of our population has some

form of dyslipidemia, whereas only 52.9% had dyslipide-
111.9
118.8
129.6
136.4
150.7
153.0
147.7

mia in the National Health and Nutrition Examination

F

Survey (NHANES) study done in USA in 2003–2006.7

(44.4)
(57.6)
(71.1)
(81.6)
(85.7)
(66.7)
(46.0)

Prevalence of high LDLC (46.0% vs 23.0%) and low

HDLC (49.6% vs 30.0%) is markedly higher with compar-
100.9
108.2
137.1
144.6
147.1
127.3
114.7
LDLC

ison to the NHANES study.

M

Although the INTERHEART study (conducted as a

(31.3)
(34.6)
(39.0)
(41.3)
(46.4)
(46.7)
(41.8)

multicentered case-control study to compare lipid param-

Lipid parameters (mean 6 SD) mg/dL

eters between patients presenting with myocardial infarc-

172.8
183.7
199.4
207.7
219.6
218.4
213.5
Total

tion (cases) and age- and sex-matched healthy controls)

observed a very high rate of low HDLC among South
(32.5)
(33.9)
(37.2)
(39.7)
(45.8)
(45.9)
(42.3)

Asians (82.3% in cases and 81% in controls) compared to

all Asians (64.3% and 59.9%),12 our study population had
176.9
184.4
196.8
206.9
226.9
227.7
222.9

a lower prevalence (49.6%); however, prevalence values

F

are still higher compared to those from NHANES

(30.1)
(35.5)
(41.5)
(44.0)
(45.0)
(44.0)
(38.5)

(29.6%). Prevalence data on high LDLC from INTER-

168.5
182.8
203.6
208.9
207.9
203.2
202.3

HEART cannot be compared because high LDLC rates

TC
M

calculated using risk stratification–based cutoffs have not

Table 4

category

been published.
19–29
30–39
40–49
50–59
60–69
,19

.69

Furthermore, data from INTERHEART show that

Age

(y)

South Asians (India, Pakistan, Bangladesh, and Sri

6 Journal of Clinical Lipidology, Vol -, No -, - 2018

Table 5 Odds ratios of risk factors in predicting different types of dyslipidemias†

Some form of
High LDLC Low HDLC High TG High TC dyslipidemia
Risk factor (n 5 2047) (n 5 2208) (n 5 1024) (n 5 896) (n 5 3445)
Advancing age 1.52* (1.21–1.67) 0.90 (0.77–0.96) 1.05 (0.84–1.21) 1.36* (1.08–1.59) 1.16* (1.01–1.29)
Female sex 1.23*(1.04–1.35) 2.94* (2.66–3.20) 0.45* (0.34–0.61) 1.30* (1.20–1.42) 1.71* (1.55–1.98)
Urban living 1.58* (1.43–1.71) 1.35* (1.20–1.49) 0.762 (0.63–0.98) 2.75* (2.52–2.89) 1.55* (1.44–1.66)
BMI 2.75* (2.48–2.95) 2.42* (2.20–2.75) 4.89* (4.32–5.44) 2.85* (2.61-3.03) 4.67* (4.21–5.07)
WC 1.21* (1.11–1.35) 1.62* (1.39–1.78) 1.23 (1.01–1.40) 1.06 (0.87–1.20) 1.48* (1.32–1.74)
Diabetes 4.48* (4.20–4.79) 1.15* (1.02–1.31) 2.05* (1.77–2.29) 0.87 (0.72–1.05) 4.25* (3.67-4.77)
Hypertension 1.92* (1.82–2.18) 0.821 (0.69–0.98) 1.50* (1.35–1.64) 1.27* (1.10–1.42) 1.79* (1.45–1.93)
Current smoking 1.66* (1.43–1.82) 1.15 (0.96–1.31) 1.36* (1.17–1.51) 0.907 (0.80–1.12) 1.37* (1.19–1.54)
Physical activity 0.96 (0.80–1.04) 0.99 (0.83–1.10) 0.75* (0.64–0.82) 1.00 (0.85–1.13) 0.68* (0.55–0.79)
BMI, body mass index; HDLC, high-density lipoprotein cholesterol; LDLC, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride;
WC, waist circumference.
Advancing age in 10-y groups, urban living compared to nonurban living (rural or estate), and current smoking compared to non–current smoking
(never smoked or not smoked within last month) were used for estimation of odds ratios.
*indicates associations of statistical significance (P , .05).
†95% confidence interval in parenthesis.

Lanka) have lower mean LDLC levels [125.2 mg/dL
(639.8) in cases and 115.4 mg/dL (637.1) in controls]
compared to non-Asians [136.2 mg/dL (642.4) in cases
and 127.1 mg/dL (639.1) in controls]. However, mean
LDLC in our study population was 135.5 mg/dL (636.6),
a value comparable to that of non-Asians in the INTER-
HEART study.
INTERHEART data also showed that South Asians have
higher TG values [163.3 mg/dL (6105.6) in cases and
169.4 mg/dL (6107.9) in controls] compared to rest of the
Asians [148.4 mg/dL (197.4) in cases and 156.7 mg/dL
(6100.3) in controls], but the values were comparable to
those of non-Asians [164.8 mg/dL (6104.8) in cases and
164.0 mg/dL (6101.8) in controls]. Interestingly, mean TG
value in our study population was much lower [121.7 mg/
dL (666.8)].
However, it is important that these data are interpreted
with caution because INTERHEART was a case-control
study that compared immediate post–myocardial infarction
patients and age- and sex-matched controls, thus probably
influencing the spectrum of population selected in terms of
age and sex. Second, lipid levels had been measured in non-
fasting samples (except those for TG level, which were
collected after an 8-hour fast). These may not be exactly
comparable with the values of 12-hour fasting samples as
chylomicrons need at least 9 hours to get totally cleared
from plasma.
Nevertheless, these comparisons suggest that Sri
Lankans have a unique pattern of dyslipidemia, charac-
terized by high LDLC, high TG, and low HDLC, for
which reasons remain poorly understood. It is possible
that complex interactions between genetic factors, socio-
cultural changes, dietary habits, and levels of physical
activity play a role. It also implies the importance of
close observation and individualization of management
strategies and need for further research into optimum
lipid management.
Determinants of dyslipidemia
Results on determinants of dyslipidemia in our study are
comparable to several others from the region. Inverse
associations between age and low HDLC, direct association
of age, and other types of dyslipidemias have been
observed in a large-scale study involving rural Chinese
adults.13 Association of hypertriglyceridemia with male sex
is also a finding seen in both these studies.
In our study population, increasing age, female sex,
living in urban sector, high BMI, central obesity, diabetes,
hypertension, insufficient physical activity, and current
smoking were associated with having some form of
dyslipidemia. Notably, this studied population had a
relatively low WC compared to other regional countries.
This is probably due to relatively young age of participants
and proportionate representation from estate and rural
sectors where central obesity is less prevalent compared
to urban setting. High BMI was the only risk factor
associated with all types of dyslipidemia, whereas central
obesity was independently associated with high LDLC and
low HDLC. Male sex was a risk factor for hypertriglycer-
idemia, whereas female sex was a risk factor for high
LDLC, low HDLC, and high TC. Advancing age was
associated with high LDLC and high TC. Low physical
activity and smoking were associated with hypertriglycer-
idemia. Hypertension was a risk factor for all types of
dyslipidemias except low HDLC. Diabetes was associated
mainly with high LDLC and hypertriglyceridemia.
Association of female sex with most types of dyslipide-
mias is a notable difference in our study. Higher prevalence
of dyslipidemia among males has been observed in several
Katulanda et al Dyslipidemia among Sri Lankan adults
other studies.14,15 Nevertheless, comparability of these
studies is limited because the definitions of dyslipidemias
are not uniform.
Association between high BMI and dyslipidemia has
been shown in several other studies. However, relationship
between central obesity and dyslipidemia varies in different
settings. In a study of rural adults in China, BMI is shown
to be associated only with hypertriglyceridemia,13 whereas
in our study, BMI relates to high LDLC and low HDLC.
Risk factor profiles for dyslipidemia in general remain
consistent across different studies. However, there are
subtle variations in specific associations probably attribut-
able partly to methodological discrepancies and partly to
unique sociocultural and lifestyle diversities that still
remain unclear.
Strengths and weaknesses
The large sample size, high response rate, and the
multistage random cluster sampling technique are major
strengths of the study. All participants fasted for 12 hours,
and testing was carried out in serum samples using most
advanced technology available in Sri Lanka with strict
adherence to procedures.
First, the exclusion of Northern and Eastern provinces
due to the ethnic conflict that prevailed in the years where
data were collected is a main drawback considering the
attempt to study a sample representative of the entire island.
Second, unavailability of data on dietary habits, which could
have been associated with dyslipidemia, prevented us from
studying this important association. Third, defining signifi-
cant family history of ischemic heart disease (IHD) and
presence of abdominal aortic aneurysms (as a risk equivalent
for IHD) were made difficult due to the unavailability of
specific data. Presence of IHD in any first-degree relative
was counted significant. As this was dependent on partici-
pant’s memory, recall bias may have been introduced.
Fourth, SLDCS, being a cross-sectional risk factor study,
carries the drawback of not being able to determine the
interactions of the risk factors longitudinally over time and
to derive conclusions on cause-effect relationships. Finally,
exclusion of participants on lipid-lowering therapy with the
aim of estimating the untreated lipid levels in our population
would have lead to underestimation of dyslipidemia preva-
lence. However, this number was relatively small (4.1%) and
unlikely to have made a significant impact.
Unavailability of other lipid parameters such as apoli-
poprotein A1 and B, small dense-LDL particles, and sub-
fractions of HDL particles and lipoprotein(a) limited our
ability to determine recently recognized risk factors for
CVD.
Conclusions
Every 3 of 4 adult Sri Lankans will have at least any one
or combination of high LDLC, low HDLC, and high TG,
7
which are risk factors for CVD. Common dyslipidemia
among both men and women is low HDLC. This pattern of
dyslipidemia is distinct and remarkable from those in other
regional and as non-Asian countries. The risk of dyslipi-
demia is associated with female sex, increasing age, high
BMI, and diabetes as expected. Common patterns associ-
ated with diabetes are high LDLC and high TG.
High prevalence, unique epidemiology and risk factor
profile of dyslipidemia among Sri Lankan adults make it
essential that more studies shall be carried out to determine
etiology and patterns, and effective screening and treatment
strategies are implemented to control the overwhelming
CVD risk faced by Sri Lankans.
Acknowledgments
SLDCS was funded by the National Science Foundation
of Sri Lanka. Support provided by the Oxford Centre for
Diabetes Endocrinology and Metabolism, UK, and the
NIHR Biomedical Research Centre Programme is grate-
fully acknowledged. Prasad Katulanda is a Commonwealth
Postgraduate Scholar. We thank the Diabetes Association of
Sri Lanka and the World Health Organization office in
Colombo for the support for lipid assays. The authors thank
all individuals and institutions who helped and worked for
the SLDCS (www.OCDEM.com/SLDCS).
Authors’ contributions: Prasad Katulanda, Rezvi Sheriff,
Godwin R. Constantine, and David R. Matthews conceived
the research question and designed the study. Prasad
Katulanda, Godwin R. Constantine, and Isurujith K.
Liyanage developed the proposal and supervised the
conduct of the study. Gaya W. Katulanda supervised the
laboratory studies and quality control. Isurujith K Liyan-
age, Harsha Anuruddhika Dissanayake, and SDN De Silva
collected data, maintained database, and conducted data
analysis. Harsha Anuruddhika Dissanayake and SDN De
Silva wrote the manuscript; and the manuscript was
critically reviewed by Prasad Katulanda, Godwin R.
Constantine, Gaya W. Katulanda, and Rezvi Sheriff. All
authors approved the final manuscript.
References
1. Modebe O. Experience with carbimazole in the drug treatment of the
hyperthyroidism of Graves’ diseases in Nigerians. East Afr Med J.
1992;69(3):153–156.
2. Third Report of the National Cholesterol Education Program (NCEP)
Expert Panel on Detection, Evaluation, and Treatment of High Blood
Cholesterol in Adults (Adult Treatment Panel III) final report. Circu-
lation. 2002;106(25):3143–3421.
3. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease
with pravastatin in men with hypercholesterolemia. West of Scotland Cor-
onary Prevention Study Group. N Engl J Med. 1995;333(20):1301–1307.
4. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute
coronary events with lovastatin in men and women with average
cholesterol levels: results of AFCAPS/TexCAPS. Air Force/Texas
Coronary Atherosclerosis Prevention Study. JAMA. 1998;279(20):
1615–1622.
8 Journal of Clinical Lipidology, Vol -, No -, - 2018
5. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H,
Kannel WB. Prediction of coronary heart disease using risk factor cat-
egories. Circulation. 1998;97(18):1837–1847.
6. Janus ED, Tideman PA, Dunbar JA, et al. Dyslipidaemia in rural
Australia: prevalence, awareness, and adherence to treatment guide-
lines in the Greater Green Triangle Risk Factor Study. Med J Aust.
2010;192(3):127–132.
7. Toth PP, Potter D, Ming EE. Prevalence of lipid abnormalities in the
United States: the National Health and Nutrition Examination Survey
2003-2006. J Clin Lipidol. 2012;6(4):325–330.
8. Thorogood M, Carter R, Benfield L, McPherson K, Mann JI. Plasma lipids
and lipoprotein cholesterol concentrations in people with different diets in
Britain. Br Med J (Clin Res Ed). 1987;295(6594):351–353.
9. Odenigbo CU, Oguejiofor OC, Odenigbo UM, Ibeh CC, Ajaero CN,
Odike MA. Prevalence of dyslipidaemia in apparently healthy profes-
sionals in Asaba, South South Nigeria. Niger J Clin Pract. 2008;11(4):
330–335.
10. Mendis S, Ekanayake EM. Prevalence of coronary heart disease
and cardiovascular risk factors in middle aged males in a defined
population in central Sri Lanka. Int J Cardiol. 1994;46(2):
135–142.
11. Katulanda P, Constantine GR, Mahesh JG, et al. Prevalence and pro-
jections of diabetes and pre-diabetes in adults in Sri Lanka–Sri Lanka
Diabetes, Cardiovascular Study (SLDCS). Diabet Med. 2008;25(9):
1062–1069.
12. Karthikeyan G, Teo KK, Islam S, et al. Lipid profile, plasma apolipo-
proteins, and risk of a first myocardial infarction among Asians: an
analysis from the INTERHEART Study. J Am Coll Cardiol. 2009;
53(3):244–253.
13. Sun GZ, Li Z, Guo L, Zhou Y, Yang HM, Suni YX. High prevalence of
dyslipidemia and associated risk factors among rural Chinese adults.
Lipids Health Dis. 2014;13(1):189.
14. Zhao WH, Zhang J, Zhai Y, et al. Blood lipid profile and prevalence of
dyslipidemia in Chinese adults. Biomed Environ Sci. 2007;20(4):
329–335.
15. Erem C, Hacihasanoglu A, Deger O, Kocak M, Topbasi M. Prevalence of
dyslipidemia and associated risk factors among Turkish adults: Trabzon
lipid study. Endocrine. 2008;34(1-3):36–51.