Prevalence of dyslipidaemia and low-density lipoprotein cholesterol target attainment: a descriptive cross-sectional study among adults in the Western Province of Sri Lanka

Abstract

Objectives

To assess the prevalence of dyslipidaemia and associated risk factors, and evaluate low-density lipoprotein cholesterol (LDL-C) target attainment among adults in the Western Province of Sri Lanka.

Design

Cross-sectional epidemiological study.

Setting

Western province, Sri Lanka.

Participants

Participants were recruited through a community-based survey of non-institutionalised adults aged ≥20 years residing in the Western Province for at least 1 year (n=1800), using multistage stratified random cluster sampling.

Primary outcome

Dyslipidaemia was defined according to the National Cholesterol Education Programme/Adult Treatment Panel III guidelines. Prevalence estimates are presented with 95% CIs. Multiple logistic regression results are reported as adjusted ORs with 95% CIs.

Secondary outcome

Cardiovascular risk in participants aged ≥40 years was assessed using the WHO laboratory-based cardiovascular disease (CVD) risk chart for South-East Asia. Achievement of LDL-C targets was evaluated according to the Sri Lankan guidelines on management for dyslipidaemia management.

Results

Data from 1333 subjects were analysed. Mean age was 49.8 (±14.9) years. The majority were females (63.6%). The age-sex standardised prevalence of any form of dyslipidaemia was 73.3% (95% CI 70.9% to 75.7%). Age standardised prevalence in females was 77.1% (95% CI 74.3% to 79.9) and males was 69.3% (95% CI 65.3% to 73.3%). The most prevalent type of dyslipidaemia was low high-density lipoprotein cholesterol (HDL-C) (46.6%, 95% CI 43.9% to 49.3%), followed by high LDL-C (32.5%, 95% CI 30.0% to 35.0%) and high triglycerides (21.7%, 95% CI 19.5% to 23.9%). Low HDL-C was positively associated with smoking (OR: 1.89, 95% CI 1.16 to 3.18) and inversely with male sex (OR: 0.29, 95% CI 19 to 0.45) and physical activity (OR: 0.71, 95% CI 0.51 to 0.99). Elevated LDL-C was associated with male sex (OR: 1.84, 95% CI 1.2 to 2.89), diabetes (OR: 5.34, 95% CI 3.53 to 8.08), and hypertension (OR: 1.62, 95% CI 1.18 to 2.23). Male sex (OR: 1.85, 95% CI 1.08 to 3.18), diabetes (OR: 1.9, 95% CI 1.4 to 2.58) and hypertension (OR: 1.81, 95% CI 1.12 to 2.91) were positively associated with elevated triglycerides, whereas urban sector (OR: 0.54, 95% CI 0.32 to 0.91) was protective. Physical activity (OR: 0.65, 95% CI 0.44 to 0.98) and male sex (OR: 0.52, 95% CI 0.31 to 0.89) inversely associated with any form of dyslipidaemia, whereas diabetes (OR: 7.08, 95% CI 3.99 to 12.55), hypertension (OR: 1.93, 95% CI 1.36 to 2.73), and body mass index (OR: 1.06, 95% CI 1.01 to 1.2) were positively associated. The majority of participants (66.6%) had a <10% 10-year CVD risk, of whom 64.9% (95% CI 60.1 to 69.8) did not achieve the LDL-C target of <3.0 mmol/L.

Conclusions

Three-fourths of adults in Western Province, Sri Lanka had any form of dyslipidaemia, more common in females. Low HDL-C was the most frequent abnormality. Most participants aged above 40 years were at low cardiovascular risk, yet two-thirds failed to meet LDL-C targets. Non-communicable disease prevention in Sri Lanka should expand through population-wide strategies, including awareness campaigns, promoting self-monitoring, targeted education and surveillance to evaluate interventions.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Large sample size, high response rate (78.2%) and use of a multistage random cluster sampling enhance representativeness and generalisability.

• Compared with Sri Lankan studies from the past decade, this study offers a more comprehensive assessment, including age-group comparisons of mean lipid levels, crude and age- and sex- standardised prevalence of dyslipidaemia subtypes, multivariable-adjusted associations, cardiovascular risk stratification using WHO risk scores, and the prevalence of achieving target lipid levels.

• Patterns of dyslipidaemia and the proportions of meeting the recommended lipid targets provide a clear epidemiological profile of cardiovascular risk in the Western Province population.

• Reducing clusters from 55 to 36 may have reduced statistical power and geographic representativeness.

• Unavailability of data on dietary habits, which could have been associated with dyslipidaemia, prevented us from studying this important association.

Introduction

Dyslipidaemia remains a major global health concern, with significant variations in prevalence across different regions. Elevated low-density lipoprotein cholesterol (LDL-C), a principal feature of dyslipidaemia, accounts for over one-third of deaths from ischaemic heart disease (IHD) and stroke.¹ According to a 2025 systematic review, Europe had the highest prevalence of high LDL-C.² Globally, 28.8% of adults have hypertriglyceridaemia, 24.1% have hypercholesterolaemia, 38.4% have low high-density lipoprotein cholesterol (HDL-C), and 18.93% have high LDL-C.² A cross-sectional survey of 35 low-income and middle-income countries (LMIC) reported that 43% and 47% of participants had high total cholesterol (TC) and high LDL-C, respectively,³ yet only 19% achieved desired lipid targets, highlighting critical gaps in detection and management.³ These patterns underscore the need for region-specific prevalence studies to inform interventions.³

Sri Lanka, a multi-ethnic LMIC with a population of 22.2 million,⁴ has undergone significant behavioural changes due to urbanisation and development. Located in southwest Sri Lanka, the Western Province is the most densely populated and urbanised, comprising Colombo, Gampaha and Kalutara districts, with 6.2 million residents.⁴ IHD is the foremost cause of mortality in Sri Lanka.⁵ Despite its clinical significance, data on dyslipidaemia in Sri Lanka remain limited. The Sri Lanka Diabetes and Cardiovascular Study (SLDCS), conducted in 2005–2006 across seven provinces, found that 77.4% of individuals had any form of dyslipidaemia, with a higher burden among females.⁶ Subsequent data on the prevalence and determinants of dyslipidaemia among community-dwelling adults were scarce until a 2019 study in Sabaragamuwa Province, which reported that 64.2% of participants had at least one abnormal lipid parameter.⁷ Furthermore, the prevalence of achieving the targeted LDL-C levels has not yet been assessed at community level. A standardised survey is thus required in Sri Lanka to generate robust prevalence estimates for cardiovascular diseases (CVDs). This would enable accurate clinical risk stratification, targeted prevention, informed clinical decision-making, evidence-based resource allocation and region-specific interventions to curb the growing burden of CVDs, including improved LDL-C target achievements. Between 2018 and 2020, we conducted a cross-sectional survey in the Western Province to estimate the prevalence of non-communicable diseases (NCDs) and to identify associated risk factors and sociodemographic determinants.⁸ The manuscript presents analyses focused on dyslipidaemia, its prevalence, clinical and demographic associations, and LDL-C target achievement in the population.

Methods

Study design and population

A cross-sectional epidemiological study was carried out from February 2018 to September 2020 in Western Province, Sri Lanka. Non-institutionalised adults aged ≥20 years who had resided in the Western Province for at least 1 year prior to data collection were recruited for this community-based study. Pregnant women were excluded.

Study sample

The multistage, stratified random cluster sampling method was used, stratifying the population by urban and rural sectors. The detailed methodology for sample size calculation, sampling and data collection has been described previously.⁸ The calculated sample size per district was Colombo (n=1068), Gampaha (n=1060) and Kalutara (n=609), totalling 2737 for Western Province.

In Sri Lanka, Grama Niladhari (GN) divisions represent the smallest administrative units and are classified as urban, rural or estate based on characteristics such as urbanisation, infrastructure and plantation activity. The required number of GN divisions per district was determined by dividing the district-specific sample size by a fixed cluster size of 50 households. Proportional allocation was then applied to assign GN divisions to each stratum based on the population of each stratum within the district. Grama Niladhari divisions were subsequently selected via simple random sampling from the district sampling frame, followed by random sampling of households from the electoral registry within each selected GN division. Adults were recruited from selected households until the target cluster size was achieved while ensuring representation of the population’s age-sex distribution. To minimise household selection bias, eligible participants were identified using a computer-assisted method similar to the Kish-grid. Although 55 GN divisions (50 households each) were required to achieve the planned sample size of 2737, financial limitations allowed data collection from only 36 GN divisions, yielding a final sample of 1800 adults while preserving proportional representation across districts.

Data collection

Data collection was made through interviewer-administered questionnaire, physical measurements and laboratory investigations of biochemical parameters. The questionnaire was adapted from the WHO STEPS survey instrument and the SLDCS^{9 10} and included sociodemographic characteristics, behavioural risk factors, history of chronic diseases and screening for NCDs. Fasting blood samples were collected for plasma glucose and lipids. Depending on proximity to the central laboratory, blood was either stored on ice and transported or serum was separated in the field. Lipid parameters, TC, HDL-C and triglycerides were measured by enzymatic photometric methods on the Cobas-C111 analyser (Roche Diagnostics, Forrennstrasse 2, Rotkreuz, Switzerland). Friedewald formula was used to calculate the LDL-C.¹¹ Glucose was measured via enzymatic oxidase colourimetry (Cobas-c111), and HbA1c by high performance liquid chromatography (Bio-Rad D–10). Blood pressure was recorded using an Omron Auto BP-HEM 7322, averaging two readings taken 15 min apart. Anthropometric measurements, including body mass index (BMI) and waist circumference (WC), were obtained. Data collection was conducted by an extensively trained team of medical graduates, nurses and science graduates, following standardised protocols. After household selection, a preliminary visit was made with the GN officer, during which one adult per household was recruited for the data collection clinic scheduled for the weekend.

Definitions

Outcomes

Lipid profiles were interpreted according to National Cholesterol Education Programme/Adult Treatment Panel III guidelines (NCEP/ATP III).¹² High TC was defined as a level >5.17 mmol/L and/or the use of lipid-lowering medications. Verification was by review of previous prescriptions. Low HDL-C was defined as <1.03 mmol/L in males and <1.29 mmol/L in females. High triglycerides were defined as >1.69 mmol/L. Specifically, LDL-C was classified as high if: (1) LDL-C was ≥2.59 mmol/L in individuals with coronary heart disease (CHD), a CHD risk equivalent, or a 10-year Framingham risk score ≥20% or (2) LDL-C was ≥3.36 mmol/L in individuals with two risk factors or a 10-year Framingham risk score <20% or (3) LDL-C was ≥4.14 mmol/L in individuals with zero to one risk factors.¹² Mixed dyslipidaemia, defined by the presence of high triglycerides, high LDL-C and low HDL-C. Cardiovascular risk assessment for participants above 40 years was based on the WHO CVD risk laboratory-based chart for South-East Asia.¹³ The achievement of LDL-C targets was assessed based on the Sri Lankan guidelines on management of dyslipidaemia.¹⁴ For primary prevention, individuals at high CVD risk (WHO risk ≥20%) should achieve an LDL-C level of <1.8 mmol/L, those at moderate risk (WHO risk 10% to 20%) should achieve <2.6 mmol/L, and those at low risk (WHO risk <10%) should achieve <3.0 mmol/L. For secondary prevention, individuals with established atherosclerotic CVD (ASCVD) should achieve an LDL-C level of <1.4 mmol/L, while those with ASCVD who experienced a second cardiovascular event within 2 years should achieve a more stringent target of 1.0 mmol/L.¹⁴

Covariates

Individuals were classified as having ‘diagnosed diabetes’ if there was unequivocal evidence of prior diagnosis, while ‘undiagnosed diabetes’ was identified based on the American Diabetes Association criteria.¹⁵ Physical activity was assessed using the short International Physical Activity Questionnaire.¹⁶ Individuals who smoked tobacco, either daily or occasionally in the past 12 months were classified as current smokers.¹⁷ Urban–rural sectors followed Sri Lankan government definitions. Hypertension was defined as 2017 American College of Cardiologists/American Heart Association guidelines.¹⁸

Data analysis

Data were collected via a tablet-based system and analysed using SPSS V.22.0 and SAS. Survey-weighted models accounted for the complex multi-stage stratified cluster sampling design, incorporating stratification, clustering and sampling weights via the STRATA, CLUSTER and WEIGHT statements in SAS. The total sampling weights were calculated as the product of the inverse probabilities of selection at each sampling stage. Age–sex-specific prevalence estimates for the Western Province were calculated using PROC SURVEYFREQ. Age- and sex-standardised prevalence estimates were derived via the direct standardisation method, with the 2018 Sri Lankan standard population (Department of Census and Statistics) as the reference. Age was categorised into 5-year groups from 20 to 59 years and ≥60 years. Prevalence estimates are reported as percentages with 95% CIs.

Normality of continuous variables was assessed using the Shapiro-Wilk test and descriptive statistics reported as means (±SD). A two-way analysis of variance assessed main effects of age group (<35, 35–44, 45–54, 55–64 and ≥65 years) and sex, as well as their interaction, on each lipid parameter. The interaction term tested whether age-related differences in lipid levels differ significantly by sex.

Univariate logistic regression assessed unadjusted associations between demographic, clinical and behavioural factors (age, sex, diabetes, hypertension, WC, BMI, physical activity, alcohol consumption and smoking) and type of dyslipidaemia (high LDL-C, low HDL-C, high triglycerides, high TC and any form of dyslipidaemia). Predictor variables for univariate logistic regression were selected a priori based on established literature, biological plausibility and data availability within the survey. BMI reflects adiposity-related inflammation influencing lipid levels.¹⁹ Alcohol consumption modulates hepatic lipid synthesis.²⁰ Smoking alters lipid homeostasis. Physical activity improves lipid profiles via enhanced lipid clearance.²¹ Potential confounders were identified using directed acyclic graphs. Age, sex, diabetes, hypertension, alcohol consumption and smoking were retained as confounders, while BMI, WC and physical activity were included as covariates. Multicollinearity was evaluated using variance inflation factors (VIFs) (<2.5 threshold).²² WC was excluded from the model due to moderate multicollinearity with BMI (VIFs: 3.18 and 3.24, respectively). The VIFs for the retained variables were all <2.5 (age: 1.30, sex: 1.62, alcohol consumption: 2.21, physical activity: 1.09, smoking: 2.07, diabetes: 1.14, hypertension: 1.21, BMI: 1.09, and sector: 1.03), indicating no multicollinearity. Multiple logistic regression model adjusted simultaneously for all retained factors, using stepwise inclusion based on a 10% change-in-estimate threshold for each variable. Residual confounding from unmeasured factors (eg, genetics) was noted as a limitation. Model adequacy was evaluated using multiple approaches, including the Hosmer-Lemeshow test for calibration, Cox & Snell and Nagelkerke pseudo-R² measures to assess explained variance, and residual and influence diagnostics (deviance and Pearson residuals, leverage and Cook’s distance) to identify potential outliers or influential observations. The area under the curve was also estimated. These results are presented in online supplemental material 1. Multiple logistic regression results are reported as adjusted ORs with 95% CIs. A two-sided p<0.05 was considered statistically significant in all statistical analysis.

Patient and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.

Results

Sociodemographic characteristics

Of the 1800 individuals invited, 1408 participated in the survey (response rate: 78.2%). Following the exclusion of 75 participants due to incomplete data, a total of 1333 participants were retained for the final analysis. Of these, 831 females (62.3%) and 485 (36.4%) from urban areas of the Western Province. Mean participant age was 49.8 (±9.1) years overall; males 49.7 (±8.93) years and females 49.8 (±8.99) years. Distribution of demographic, behavioural and clinical characteristics is presented in table 1.

Table 1. Demographic, behavioural and clinical characteristics of the study participants (N=1333).

Characteristic	Male (n=502)	Female (n=831)	Total (n=1333)
Sector of living*
Urban	32.6 (13.5 to 51.7)	40.9 (20.5 to 61.4)	37.8 (18.2 to 57.5)
Rural	67.4 (48.3 to 86.5)	59.0 (38.6 to 79.5)	62.2 (42.5 to 81.8)
Diabetes mellitus*	41.9 (34.7 to 49.3)	33.3 (27.6 to 39.0)	36.6 (31.8 to 41.4)
Hypertension*	55.9 (50.8 to 61.0)	56.5 (51.6 to 61.4)	56.3 (52.1 to 60.5)
Ischaemic heart disease*	4.7 (2.6 to 6.9)	4.3 (2.6 to 5.9)	4.5 (3.1 to 5.8)
Current smokers* (smoking for last year)	42.8 (31.4 to 54.3)	1.2 (0.3 to 1.9)	16.7 (11.9 to 21.5)
Alcohol consumption* (within last 6 months)	56.8 (46.2 to 67.4)	6.8 (2.8 to 10.9)	25.5 (18.9 to 32.1)
Physical activity*
Insufficient physical activity	49.2 (41.8 to 56.6)	59.3 (52.2 to 66.3)	55.5 (48.8 to 62.2)
Moderately active	37.9 (32.9 to 43.0)	38.1 (31.1 to 45.0)	38.0 (32.6 to 43.4)
Highly active	12.8 (8.7 to 16.9)	2.7 (1.2 to 4.2)	6.5 (4.4 to 8.5)
Body mass index (kg/m2)†	24.7 (±4.59)	25.9 (±5.13)	25.7 (±4.9)
Waist circumference (cm)†	88.3 (±11.7)	87.6 (±11.5)	87.9 (±11.6)
Waist-to-height ratio†	0.54 (±0.06)	0.57 (±0.07)	0.56 (±0.08)
TC (mmol/L)†	5.15 (±1.13)	5.13 (±1.12)	5.14 (±1.12)
HDL-C (mmol/L)†	1.23 (±0.38)	1.3 (±0.36)	1.28 (±0.39)
LDL-C (mmol/L)†	3.22 (±0.91)	3.23 (±0.97)	3.23 (±0.94)
Triglycerides (mmol/L)†	1.46 (±0.39)	1.29 (±0.35)	1.35 (±0.42)

^*Values are presented as percentage and 95% CI.

^†Values are presented as the mean and SD.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol.

Prevalence of dyslipidaemia

The crude prevalence of low HDL-C was 48.9% (95% CI 42.3% to 55.4%), high LDL-C was 39.6% (95% CI 36.7% to 42.5%) and high triglycerides was 20.4% (95% CI 15.9% to 24.9%). The crude prevalence of high TC was found to be 42.3% (95% CI 36.5% to 48.2%) and having any form of dyslipidaemia was 79.5% (95% CI 75.5% to 83.4%). Mixed dyslipidaemia was observed in 6.4% (95% CI 4.6% to 8.2%). Distribution of different types of dyslipidaemia across sociodemographic characteristics is presented in table 2. Figure 1 illustrates isolated and overlapping forms of dyslipidaemia in the study population. The age-standardised and sex-standardised prevalence of different types of dyslipidaemia is presented in table 3.

Table 2. Prevalence of dyslipidaemia across the sociodemographic characteristics (%, 95% CI).

	High LDL	Low HDL	High triglycerides	Any form of dyslipidaemia
District
Colombo	43.1 (38.5 to 47.8)	57.8 (47.8 to 67.7)	16.4 (11.1 to 21.7)	84.4 (78.6 to 90.2)
Gampaha	35.8 (32.4 to 39.3)	42.9 (35.2 to 50.7)	26.8 (21.4 to 32.2)	77.6 (71.8 to 83.4)
Kalutara	40.7 (35.4 to 46.1)	41.6 (29.9 to 53.3)	13.3 (4.9 to 21.7)	71.7 (66.4 to 76.9)
Sector
Urban	40.8 (35.7 to 45.9)	56.3 (45.4 to 67.2)	15.1 (10.2 to 19.9)	83.4 (77.2 to 89.7)
Rural	38.9 (35.6 to 42.1)	44.3 (37.8 to 50.9)	23.6 (18.2 to 29.0)	77.1 (72.4 to 81.8)
Education
Degree and above	22.9 (11.2 to 34.7)	28.8 (18.2 to 39.3)	9.1 (1.9 to 16.3)	47.7 (35.4 to 59.9)
Tertiary	33.4 (26.7 to 40.1)	47.6 (38.1 to 57.1)	20.1 (13.9 to 26.4)	77.2 (71.5 to 82.9)
Secondary	41.9 (37.4 to 46.4)	51.0 (44.7 to 57.3)	21.4 (16.2 to 26.5)	82.5 (78.1 to 86.9)
Primary	54.9 (45.9 to 63.9)	49.4 (34.2 to 64.6)	20.9 (9.3 to 32.5)	84.2 (78.5 to 89.9)
Monthly family income (US$)
>400	28.6 (21.0 to 36.1)	35.6 (25.6 to 45.6)	25.5 (26.5 to 34.4)	65.2 (52.2 to 78.2)
200–400	41.8 (37.7 to 45.9)	51.4 (44.4 to 58.3)	17.8 (13.0 to 22.5)	81.1 (77.4 to 84.7)
<200	38.3 (31.1 to 45.6)	47.8 (37.6 to 58.0)	25.3 (14.8 to 35.8)	81.2 (74.9 to 87.4)
Occupation
Managers and professionals	31.3 (21.8 to 40.8)	44.7 (29.2 to 60.2)	13.9 (4.5 to 23.4)	68.5 (58.4 to 78.5)
Technical and clerical	36.7 (26.5 to 46.8)	38.5 (26.2 to 50.8)	20.1 (11.3 to 28.9)	72.7 (58.4 to 87.9)
Self-employers/farmers/skilled workers	39.9 (32.4 to 47.4)	39.1 (29.9 to 48.2)	23.6 (13.9 to 33.2)	71.8 (63.5 to 80.0)
Elementary workers	48.9 (38.8 to 59.1)	43.3 (32.9 to 53.6)	29.0 (19.9 to 38.2)	78.7 (69.4 to 88.0)
Retired	49.0 (37.1 to 60.9)	47.0 (34.7 to 59.4)	29.7 (21.1 to 38.3)	90.0 (82.9 to 97.1)
Unemployed	38.2 (33.6 to 42.9)	54.7 (37.7 to 52.9)	17.8 (11.7 to 23.9)	82.6 (77.4 to 87.8)
Total	39.6 (36.7 to 42.5)	48.9 (42.3 to 59.0)	20.4 (15.9 to 24.9)	79.5 (75.5 to 83.4)

HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Figure 1. Overlap and isolated types of dyslipidaemia (Venn diagram). HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglycerides.

Table 3. Prevalence of dyslipidaemia types; overall age- and sex- standardised and age-standardised separately for males and females (%, 95% CI).

Types of dyslipidaemia	Age- and sex-standardised prevalence	Age-standardised prevalence in males	Age-standardised prevalence in females
High LDL-C	32.5 (30.0 to 35.0)	38.7 (34.4 to 42.9)	26.8 (23.7 to 29.8)
Low HDL-C	46.6 (43.9 to 49.3)	29.8 (25.8 to 33.8)	62.3 (58.9 to 65.6)
High triglycerides	21.7 (19.5 to 23.9)	31.2 (27.1 to 35.2)	12.6 (10.6 to 15.1)
High TC	42.0 (39.4 to 44.7)	44.9 (40.5 to 49.2)	39.3 (36.0 to 42.7)
Mixed dyslipidaemia	5.9 (4.7 to 7.2)	7.1 (4.9 to 9.4)	4.9 (3.4 to 6.4)
Any form of dyslipidaemia	73.3 (70.9 to 75.7)	69.3 (65.3 to 73.3)	77.1 (74.3 to 79.9)
Already diagnosed dyslipidaemia	16.0 (14.1 to 17.9)	14.9 (11.8 to 17.9)	17.1 (14.5 to 19.7)
Newly diagnosed dyslipidaemia	57.3 (54.7 to 59.9)	54.5 (50.1 to 58.8)	60.0 (56.7 to 63.3)

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol.

Figure 2 illustrates the distribution of abnormal dyslipidaemia types across different age groups and sex. The prevalence of elevated triglycerides peaked in males aged 35–44 years and in females aged over 65 years. Across all age groups, females consistently had higher HDL-C levels compared with males. The prevalence of low HDL-C reached its highest level in males aged over 65 years and in females aged less than 35 years. The prevalence of elevated LDL-C was highest in over 65 years in both males and females.

Figure 2. Age-specific and sex-specific prevalence of (a) low HDL-C, (b) high LDL-C and (c) high triglycerides. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.

Figure 3 summarises the mean lipids levels in males and females across age groups. Mean HDL-C levels (figure 3a) did not differ significantly across age groups (p=0.06) and were higher in one sex compared with the other (p=0.003). There was no significant sex-by-age group interaction (p=0.15). Mean LDL-C levels (figure 3b) varied modestly across age groups (p=0.04) and showed no difference between sexes (p=0.74), although a significant age-by-sex interaction was observed (p<0.001). Triglyceride levels (figure 3c) differed across age groups (p=0.004) and between sexes (p<0.001), with a significant sex-by-age group interaction (p<0.001). TC levels (figure 3d) varied modestly across age groups (p=0.03) and showed no difference between sexes (p=0.8), accompanied by a significant sex-by-age group interaction (p<0.001).

Figure 3. Mean values of (a) HDL-C, (b) LDL-C, (c) triglycerides and (d) total cholesterol (TC) in males and females across age groups. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.

Results of the univariate logistic regression are presented as unadjusted ORs with 95% CI in table 4.

Table 4. Unadjusted ORs with 95% CIs for risk factors associated with different types of dyslipidaemia (n=1333).

Risk factor	Low HDL-C	High LDL-C	High TG	High TC	Any form of dyslipidaemia
Age	0.98 (0.97 to 0.99)*	1.05 (1.04 to 1.06)*	1.01 (0.99 to 1.02)	1.001 (0.99 to 1.01)	1.04 (1.03 to 1.05)*
Sex
Male	0.29 (0.22 to 0.4)*	1.78 (1.32 to 2.40)*	2.16 (1.56 to 2.99)*	1.06 (0.79 to 1.42)	0.79 (0.53 to 1.19)
Female	Ref
Sector
Urban	1.62 (0.96 to 2.72)	1.09 (0.84 to 1.39)	0.58 (0.35 to 0.94)*	0.73 (0.44 to 1.20)	1.49 (0.89 to 2.53)
Rural	Ref
Diabetes versus non-diabetes	1.01 (0.73 to 1.39)	7.16 (4.93 to 10.38)*	2.09 (1.49 to 2.95)*	0.82 (0.63 to 1.07)	9.05 (5.26 to 15.56)*
Hypertension versus non-hypertensive	0.90 (0.63 to 1.29)	2.58 (1.74 to 3.82)*	2.15 (1.56 to 2.97)*	1.03 (0.77 to 1.39)	3.01 (2.07 to 4.39)*
Body mass index	1.06 (1.03 to 1.09)*	1.004 (0.98 to 1.03)	1.04 (1.02 to 1.08)*	0.98 (0.96 to 1.02)	1.08 (1.04 to 1.12)*
Waist circumference (cm)	1.01 (0.99 to 1.02)	1.02 (1.01 to 1.03)*	1.03 (1.01 to 1.05)*	0.99 (0.98 to 1.01)	1.04 (1.02 to 1.05)*
Physical activity versus insufficiently active	0.72 (0.56 to 0.93)*	0.74 (0.58 to 0.95)*	1.01 (0.66 to 1.55)	1.37 (0.96 to 1.96)	0.58 (0.40 to 0.82)*
Current alcohol consumers versus non-consumers	0.44 (0.31 to 0.63)*	1.29 (0.95 to 1.76)	2.05 (1.59 to 2.62)*	0.89 (0.59 to 1.32)	1.17 (0.88 to 1.55)
Current smoking versus non-smoking	0.51 (0.34 to 0.78)*	1.83 (1.29 to 2.59)*	2.06 (1.52 to 2.79)*	0.87 (0.65 to 1.15)	1.52 (0.88 to 2.64)

^*Significant associations at p<0.05.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides.

Multiple logistic regression analysis identified several factors associated with lipid parameters and presented as adjusted ORs with 95% CI in table 5. Smoking was associated with higher odds of low HDL-C (OR: 1.89, 95% CI 1.16 to 3.18). Each 1-year increase in age was associated with a 2% reduction in the odds of low HDL-C (OR: 0.98, 95% CI 0.97 to 0.99), while male sex (OR: 0.29, 95% CI 0.19 to 0.45) and physical activity (OR: 0.71, 95% CI 0.51 to 0.99) were also associated with lower odds of low HDL-C. Each additional year of age was associated with increased odds of elevated LDL-C (OR: 1.03, 95% CI 1.02 to 1.05). Male sex (OR: 1.84, 95% CI 1.20 to 2.89), diabetes (OR: 5.34, 95% CI 3.53 to 8.08), and hypertension (OR: 1.62, 95% CI 1.18 to 2.23) were also associated with higher odds of elevated LDL-C. An increase of one BMI unit was associated with increased odds of elevated triglyceride levels (OR: 1.05, 95% CI 1.01 to 1.09). Male sex (OR: 1.85, 95% CI 1.08 to 3.18), diabetes (OR: 1.90, 95% CI 1.40 to 2.58) and hypertension (OR: 1.81, 95% CI 1.12 to 2.91) were similarly associated with increased odds of elevated triglycerides. In contrast, residence in urban areas was associated with lower odds of elevated triglycerides compared with rural residence (OR: 0.54, 95% CI 0.32 to 0.91). Each one-unit increase in BMI was associated with a 6% increase in the odds of having any form of dyslipidaemia (OR: 1.06, 95% CI 1.01 to 1.20). Physical activity (OR: 0.65, 95% CI 0.44 to 0.98) and male sex (OR: 0.52, 95% CI 0.31 to 0.89) were associated with lower odds of any form of dyslipidaemia, whereas diabetes (OR: 7.08, 95% CI 3.99 to 12.55) and hypertension (OR: 1.93, 95% CI 1.36 to 2.73) were associated with higher odds.

Table 5. Adjusted ORs with 95% CIs for risk factors associated with different types of dyslipidaemia (n=1333).

Risk factor	Low HDL-C	High LDL-C	High TG	High TC	Any form of dyslipidaemia
Age	0.98* (0.97 to 0.99)	1.03* (1.02 to 1.05)	0.99 (0.97 to 1.02)	1.01 (0.99 to 1.02)	1.01 (0.99 to 1.03)
Male versus female	0.29* (0.19 to 0.45)	1.84* (1.2 to 2.89)	1.85* (1.08 to 3.18)	1.22 (0.82 to 1.91)	0.52* (0.31 to 0.89)
Urban living versus Rural	1.48 (0.84 to 2.53)	0.92 (0.64 to 1.31)	0.54* (0.32 to 0.91)	0.75 (0.45 to 1.26)	1.21 (0.72 to 2.05)
Diabetes versus non-diabetes	1.29 (0.98 to 1.75)	5.34* (3.53 to 8.08)	1.90* (1.40 to 2.58)	0.81 (0.61 to 1.05)	7.08* (3.99 to 12.55)
Hypertension versus non-hypertensive	0.96 (0.66 to 1.42)	1.62* (1.18 to 2.23)	1.81* (1.12 to 2.91)	1.06 (0.72 to 1.56)	1.93* (1.36 to 2.73)
Body mass index	1.04 (0.99 to 1.13)	1.02 (0.95 to 1.09)	1.05* (1.01 to 1.09)	0.99 (0.97 to 1.02)	1.06* (1.01 to 1.2)
Physical activity versus insufficiently active	0.71* (0.51 to 0.99)	0.94 (0.62 to 1.41)	0.89 (0.63 to 1.29)	1.43 (0.99 to 2.07)	0.65* (0.44 to 0.98)
Current smoking versus non-smoking	1.89* (1.16 to 3.18)	1.91 (0.94 to 3.89)	1.14 (0.59 to 2.18)	0.8 (0.39 to 1.59)	2.35 (0.89 to 6.19)
Current alcohol consumers versus non-consumers	0.73 (0.45 to 1.19)	0.55 (0.25 to 1.24)	1.32 (0.71 to 2.45)	0.81 (0.41 to 1.59)	1.13 (0.57 to 2.24)

^*Significant associations at p<0.05.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides.

LDL-C target attainment

Among the 912 participants aged ≥40 years, 41.6% (95% CI 36.6% to 46.6%) were classified into the <5% 10-year cardiovascular CVD risk category, 25.0% (95% CI 20.2% to 29.9%) into the 5%–10% category, 27.7% (95% CI 24.3% to 31.0%) into the 10%–20% category, 4.7% (95% CI 3.0% to 6.3%) into the 20%–30% category, and 0.9% (95% CI 0.3% to 2.7%) into the ≥30% category. Figure 4 depicts the proportion of individuals who did not achieve recommended LDL-C targets across CVD risk categories, according to national dyslipidaemia management guidelines.¹⁴ Among participants with established ASCVD, who comprised 6.2% of the study population, the vast majority (95.1%; 95% CI 87.9% to 100.0%) failed to achieve the recommended LDL-C target of <1.4 mmol/L. Notably, no recurrent ASCVD events were documented during the 2-year period. Among individuals classified as very high risk based on a ≥20% estimated 10 year CVD risk (5.7% of participants), 92.9% (95% CI 81.9% to 100.0%) did not meet the LDL-C target of <1.8 mmol/L. In the moderate-risk group with a 10%–20% estimated 10-year CVD risk, representing 27.7% of the cohort, 67.5% (95% CI 59.2% to 75.8%) failed to attain the LDL-C target of <2.6 mmol/L. The largest proportion of participants (66.6%) were classified as low risk, with an estimated 10-year CVD risk of <10%. Among this group, nearly two-thirds (64.9%; 95% CI 60.1% to 69.8%) did not achieve the recommended LDL-C target of <3.0 mmol/L. Additionally, 23.5% of all participants reported a prior diagnosis of dyslipidaemia, of whom 79.4% (95% CI 72.3% to 86.5%) were receiving lipid-lowering therapy.

Figure 4. Proportion of participants achieving LDL-C targets across different cardiovascular disease risk groups. ASCVD, atherosclerotic cardiovascular disease; LDL-C, low-density lipoprotein cholesterol.

Discussion

The survey represents the latest comprehensive assessment of dyslipidaemia among adults ≥20 years in Western Province, Sri Lanka conducted across 36 Grama Niladhari divisions. In this article, we describe the prevalence of dyslipidaemia, its associations and the achievement of LDL-C targets in a representative population of the Western province in Sri Lanka. The crude prevalence among the adults in the Western province of Sri Lanka was 48.9% for low HDL-C, 39.6% for high LDL-C, 20.4% for high triglycerides and 42.3% for high TC. Overall, 79.5% had at least one form of dyslipidaemia, and 6.4% had mixed dyslipidaemia. Only a limited number of surveys have been conducted in Sri Lanka, and most were confined to specific geographical areas. The SLDCS, our previous national survey conducted in 2005–2006 represents the earliest community-based assessment of dyslipidaemia in Sri Lanka. National crude prevalence rates from the SLDCS (using NCEP/ATP III criteria)¹² were 49.6% for low HDL-C, 23.0% for high triglycerides, 46.0% for high LDL-C, 77.4% for any dyslipidaemia, and 7.6% for mixed dyslipidaemia.⁶ The current Western Province survey demonstrates lower prevalence across most parameters even after 15 years. However, the absence of province-specific data in the SLDCS, which limits precise regional benchmarking, precludes direct comparisons and constitutes a key limitation of that study. This observed reduction likely reflects Sri Lanka’s strengthened primary healthcare system, NCD policies²³ and lipid-lowering medication use (22.4%, 95% CI 18.3% to 26.5%). Similar downward trends in dyslipidaemia parameters are evident globally. United States NHANES data showed reductions in high TC and low HDL-C from 2007 to 2018,²⁴ attributable to trans-fat regulations.²⁵ In Europe, population-level declines in TC and non-HDL-C have occurred alongside rising HDL-C.²⁶ Additionally, the age-standardised mortality attributable to high LDL-C decreased by 64.9% from 1990 to 2019.²⁷

An adolescent survey conducted in 2006 in the Kandy municipal council area reported high triglycerides (males: 6.1%, females: 4.5%), high LDL-C (males: 26.8%, females: 31.2%), and low HDL-C (males: 8.4%, females: 6.8%).²⁸ These findings indicate a relatively high prevalence compared with expectations for this age group, potentially attributable to factors such as the urbanised lifestyle increasingly adopted by school-going adolescents. When compared with a prior 1997 study assessing dyslipidaemia among adolescents in urban areas of Colombo, Negombo and Kurunegala, which reported high LDL-C in 21.9% and low HDL-C in 27.3%,²⁹ the results of the Kandy survey are comparable. Findings of these surveys align predictively with our adult data, indicating the persistent burden of dyslipidaemia across generations and underscoring the need for school-based NCD interventions. A recent study from Sri Lanka’s Northern Province assessed lipid profiles among apparently healthy adults recruited as controls for another study, all of whom had no prior diagnoses or abnormalities on NCD screening. They defined the desirable lipid levels as TC<5.17 mmol/L, LDL-C<3.36 mmol/L, TG<1.69 mmol/L and HDL-C≥1.03 mmol/L for men and ≥1.29 mmol/L for women. High LDL-C was in 35.8% of participants, high triglycerides in 22.8% and low HDL-C in 31.6%.³⁰ Compared with the Northern Province, the Western Province demonstrated a higher prevalence of low HDL-C despite similar triglyceride and LDL-C levels, likely reflecting lifestyle differences. Urban residents, particularly in Colombo, tend to have more sedentary occupations, prolonged sitting and lower physical activity, all of which are known to lower HDL-C without strongly affecting triglycerides or LDL-C. Exercise-activated lipoprotein lipase (LPL) hydrolyses triglycerides in chylomicrons and very low density lipoprotein, promoting HDL-C formation, whereas physical inactivity suppresses this process and reduces HDL-C levels.³¹ Reverse cholesterol transport, whereby HDL removes excess cholesterol for hepatic excretion, is also exercise dependent and impaired by inactivity through altered hepatic lipase regulation.³² Furthermore, higher consumption of refined carbs, fast foods and trans fats in urban Western diets can suppress HDL-C more than in the relatively traditional, rice-and-vegetable-based Northern cuisine. A contemporaneous study from Colombo district reported a hypercholesterolaemia prevalence of 31.9% (TC>6.21 mmol/L),³³ while our study observed a higher prevalence of elevated TC (38.3%) using a lower threshold (TC>5.17 mmol/L). A provincial survey from the predominantly rural Sabaragamuwa Province (2019–2020) reported an overall dyslipidaemia prevalence of 64.2%, with high LDL-C (37.7%), high triglycerides (21.3%), and low HDL-C (29.0%).⁷ In contrast, our study found low HDL-C to be the predominant abnormality, followed by high LDL-C and high triglycerides. Rural diets rich in coconut or palm saturated fats are likely to drive LDL elevation despite limited medication access, while urban sedentary behaviour, refined carbohydrate intake and stress selectively suppress HDL-C through impaired LPL activity and reverse cholesterol transport.³⁴ While the national NCD Risk Factor Survey reported a dyslipidaemia prevalence of 15.2%, nearly half of adults had never undergone lipid testing.³⁵ In comparison, diagnosed dyslipidaemia in our study (23.5%) was comparable to that reported from Sabaragamuwa Province (21.5%).⁷

The ICMR-INDIAB national cross-sectional study conducted between 2008 and 2020 reported prevalence of 24.0% for hypercholesterolaemia, 32.1% for hypertriglyceridaemia, 66.9% for low HDL-C, 20.9% for high LDL-C and 81.2% for any type of dyslipidaemia.³⁶ Dyslipidaemia was defined using the NCEP/ATP III.¹² The prevalence of any form of dyslipidaemia was comparable to that observed in our study (79.5%). However, the prevalence of low HDL-C and high triglycerides in ICMR-INDIAB was higher than those observed in our population (48.9% and 20.4%, respectively), while the prevalence of high LDL-C was lower than ours (39.6%). The ICMR-INDIAB study was a multiphase, cross-sectional survey conducted over several phases, capturing evolving national trends such as increasing urbanisation and dietary shifts toward processed foods, which may have contributed to the higher prevalence of hypertriglyceridaemia (32.1%) and low HDL-C (66.9%) in urban-dominant phases. The first phase of the study reported prevalence of 13.9% for hypercholesterolaemia, 29.5% for hypertriglyceridaemia, 72.3% for low HDL-C, 11.8% for high LDL-C, and 79.0% for any type of dyslipidaemia.³⁷ Despite being cross-sectional overall, pooling data across phases may have averaged out temporal changes, making national estimates less comparable with a uniform contemporary study such as ours. In this context, higher prevalence of high LDL-C (39.6%) observed in our study may reflect more recent dietary patterns that were not captured in the earlier phases of ICMR-INDIAB. This highlights the value of contemporaneous benchmarking. A cross-sectional survey in the UAE reported overall dyslipidaemia prevalence of 72.5% using American College of Cardiology/American Heart Association guidelines, including 42.8% high TC, 29% high triglycerides, 42.5% low HDL-C and 38.6% high LDL-C.³⁸ Our study shows comparable rates across these parameters, despite methodological differences such as variations in guidelines. Conducted 6 to 7 years before ours, this earlier UAE survey reflects stable cardiometabolic risks in similar demographics, potentially driven by shared factors like urbanisation, high-carb diets and genetic predispositions common to Arab populations. The close match reinforces the persistence of dyslipidaemia patterns over time, underscoring the need for ongoing surveillance. Minor variations in higher triglyceride prevalence may stem from evolving lifestyle interventions post-2013. In a community-based cross-sectional study conducted from August one to November 30, 2020, in Shangcheng District, an urban centre in Hangzhou, the capital of Zhejiang Province, China, researchers reported notably lower dyslipidaemia prevalence rates than those observed in our study. Specifically, they found high TC in 17.8% of participants, high triglycerides in 12.85%, low HDL-C in 10.96%, high LDL-C in 3.91%, and overall dyslipidaemia in 35.96%. Dyslipidaemia was defined as TC≥6.2 mmol/L and/or triglycerides≥2.3 mmol/L and/or HDL-C≤1.0 mmol/L and/or LDL-C≥4.1 mmol/L, or current use of lipid-lowering agents, or a prior diagnosis.³⁹ These estimates stand in stark contrast to our higher prevalence figures, even though Shangcheng represents a highly urbanised population and data collection occurred over a similar timeframe. However, differences in diagnostic definitions underscore the inherent challenges in making reliable cross-study comparisons, complicating efforts to interpret true population-level data. Beyond Asian contexts, a 2011 community-based study in rural Uganda reported low HDL as the most prevalent lipid abnormality, affecting 71.3% (70.2%–72.3%) of participants, followed by high TC at 6.0%, high LDL-C at 5.2%, and high triglycerides at 5.0%.⁴⁰ Dyslipidaemia is aligned with NCEP/ATP III guidelines,¹² mirroring our definitions. Our rural cohort exhibited even higher rates across all measures, high LDL-C (38.9%), low HDL-C (44.3%), high triglycerides (23.6%) and overall dyslipidaemia (77.1%) (table 2). The particularly elevated prevalence of low HDL-C, outpacing other parameters, likely stems from rural-specific factors prevalent in both Ugandan and our settings, including low physical activity, diets rich in carbohydrates and poor in healthy fats, tobacco use, genetic predispositions in certain populations and limited access to healthcare for modifiable risks like obesity. This pattern underscores how low HDL-C often emerges as the dominant dyslipidaemia marker in resource-limited rural areas worldwide, signalling a need for targeted interventions focused on lifestyle and affordable HDL-boosting strategies. Nationwide longitudinal studies in Sri Lanka are essential to explore the relationship between low HDL-C and body composition, particularly in the context of the country’s rising burden of cardiovascular and metabolic diseases. East Asia experienced a near tripling of IHD and stroke deaths attributable to high non-HDL-C (from 250 000 to 860 000) between 1990 and 2017, with Southeast Asia showing a similar trend (110 000 to 310 000).⁴¹

While our study establishes critical dyslipidaemia benchmarks for Sri Lanka, its comparisons with international surveys deliver a globally pertinent message: high prevalence patterns driven by urbanisation, dietary shifts and rural lifestyle gaps mirror those observed in India (81.2%), the UAE (72.5%) and rural Uganda, signalling persistent cardiometabolic risks in urbanising LMIC populations. In contrast, findings from urban China underscore the need for diagnostic harmonisation, emphasising contemporaneous NCEP/ATP III–based studies for reliable global benchmarking. These insights advocate targeted interventions such as HDL-boosting strategies and longitudinal surveillance, positioning Sri Lanka as a potential model for curbing CVD surges seen in Southeast Asia.

In our study, male sex and physical activity were associated with lower odds of any form of dyslipidaemia, while BMI, diabetes and hypertension increased the risk. Age showed no association (OR=1.01; 95% CI 0.99 to 1.03). Conversely, the Sabaragamuwa study reported links with age and BMI, and not with hypertension, diabetes or male sex.⁷ Similar to our findings, SLDCS revealed positive associations between various forms of dyslipidaemia and advancing age, female sex, physical inactivity, hypertension and higher BMI.⁶ Diabetes and urban living also were contributing factors for any forms of dyslipidaemia in SLDCS.⁶ Even though urban living was positively associated with high LDL-C, low HDL-C and high TC in SLDCS,⁶ urban living was negatively associated with high triglycerides in our findings. These differences may be attributed to increasing urbanisation and highlight the need to reconsider the urban–rural classification within Sri Lanka. We acknowledge that Sri Lanka’s 1987 urban–rural classification may not fully reflect current urbanisation patterns, potentially affecting urban–rural comparisons. However, this classification remains the standard in national datasets, and therefore the urban–rural differences should be interpreted cautiously. Notably, the Sabaragamuwa study did not account for urban–rural stratification.⁷ Smoking and alcohol consumption were not associated with dyslipidaemia in residents of Sabaragamuwa province,⁷ while smoking was positively associated with low HDL (OR=1.89; 95% CI 1.16 to 3.18) in our study. In contrast, the SLDCS reported that current smoking was associated with high LDL-C, high triglycerides and any form of dyslipidaemia.⁶ However, it was noted that the percentage of smokers decreased from 18.3%⁴² in 2005 to 9.1%⁴³ in 2020, which could explain the varying associations with different forms of dyslipidaemia in our findings.

Strengths

The large sample size, high response rate (78.2%), and use of a multistage random cluster sampling design are major strengths of this study, enhancing its representativeness and generalisability. All participants underwent a 12-hour fasting period, and serum lipid analyses were performed using the standard laboratory technology with strict adherence to quality-controlled procedures. The use of standardised protocols for data collection, field-based sample processing, centralised laboratory analysis and prevalence estimation ensured the generation of robust and contemporary data, facilitating meaningful national and international comparisons.

Few community-based surveys have examined the prevalence of dyslipidaemia among Sri Lankan adults. Compared with studies published in the past decade, this study provides a more comprehensive assessment, including the demographic distribution of the sample, mean lipid values and age-group comparisons, crude prevalence of different types of dyslipidaemia by age, sex and demographic characteristics, overlap and isolated forms of dyslipidaemia, age-standardised and sex-standardised prevalence of dyslipidaemia subtypes, univariate associations with clinical risk factors and multivariable-adjusted associations accounting for potential confounders. In addition, we assessed cardiovascular risk stratification using the WHO risk score and estimated the proportion of participants who achieved target lipid levels according to their cardiovascular risk category. To the best of our knowledge, no previous community-based study in Sri Lanka has presented such a comprehensive and integrated description of dyslipidaemia and its cardiovascular risk implications. By characterising patterns of dyslipidaemia, a major risk factor for CVD, and evaluating the proportion of the community meeting recommended lipid targets, this study provides a clear epidemiological profile of cardiovascular risk of the Western Province population, the most urbanised region of Sri Lanka, where access to healthcare services is relatively optimal.

Limitations

We acknowledge that reducing clusters from 55 to 36 due to COVID-19 and financial constraints may have affected statistical power and subregional representativeness. Data collection was completed while maintaining the proportional representation of strata in GN divisions within each district. Furthermore, we conducted a post hoc assessment that estimated the intracluster correlation coefficient for the primary outcome at 0.02. This corresponds to a design effect of 1.98 and an effective sample size of approximately 909. The weighted prevalence estimates generated using PROC SURVEYFREQ retained acceptable precision, with only a modest widening of CIs compared with the originally planned 55-cluster design. Nevertheless, these slightly wider intervals likely reflect reduced between-cluster variability rather than true population differences. We also approximated the statistical power post-hoc for detecting differences between two equal subgroups. With 55 clusters, power was approximately 48% for a 5%-point difference, 87% for 8%, and 97% for 10%. With 36 clusters, power decreased to about 34%, 70% and 88% for the same differences. This indicates that the reduction mainly affects detection of smaller differences, while larger differences remain detectable with reasonable power. Since the primary aim of our study is to estimate prevalence, our focus is on precision rather than hypothesis testing. Although the overall response rate of 78.2% is generally considered acceptable for community survey research, the response proportions differed significantly across age groups and sex, based on a cluster-adjusted chi-square test (p<0.05). This heterogeneity indicates the potential for selection bias, which may affect the representativeness of the sample and limit the generalisability of findings, particularly for subgroups with lower participation. Future studies should consider additional follow-up attempts, extended data-collection periods or participant incentives to improve response rates and minimise differential non-response. Approximately 75 cases (~5% of the total 1,408) were excluded at the data cleaning due to incomplete key variables such as age, sex, laboratory results and blood pressure values. This missingness is primarily attributable to random data entry omissions and is assumed to be Missing Completely at Random (MCAR). Under this assumption, complete-case analysis is expected to provide unbiased estimates. The exclusion of this relatively small proportion of data likely has minimal impact on statistical power or precision, with a final analytic sample size of 1333 participants. However, we acknowledge the limitation of not performing sensitivity analyses to evaluate the robustness of findings under alternative missing data mechanisms. While imputation is not strictly necessary under MCAR, it remains a valuable option for improving efficiency and precision.

Conclusions

Three-fourths of adults in Western Province, Sri Lanka had any form of dyslipidaemia, more common in females. Low HDL-C was the most common lipid abnormality. High BMI, hypertension, diabetes and smoking were high risk factors for dyslipidaemia. While most participants aged≥40 years were at low cardiovascular risk, two-thirds failed to meet recommended LDL-C targets, rising to 95% among those with established CVD. These findings underscore the need for population-wide strategies, including mass awareness campaigns, promotion of regular self-monitoring, workplace programmes, and targeted health education for lower socioeconomic groups, alongside primordial prevention and robust surveillance to track trends and evaluate interventions.

Data availability statement

Data are available on reasonable request.