Rural–urban differences in lipid abnormalities among middle-aged and older Indians

Abstract

Background

Dyslipidemia is a major risk factor for cardiovascular diseases (CVD). The prevalence of dyslipidemia varies by geographic location, often higher in urban populations. Our study aimed to assess the prevalence of dyslipidemia in the aging Indian population and compare the rural–urban differences in lipid abnormalities.

Methods

We analyzed baseline cross-sectional data from two longitudinal aging cohorts in rural and urban southern India- 2,797 participants from the rural (CBR-SANSCOG) cohort in Srinivaspura, Karnataka, and, 430 participants from the urban (CBR-TLSA) cohort in Bangalore, Karnataka. Participants aged ≥ 45 years were included, and those with dementia, severe psychiatric/medical illnesses, and severe visual/hearing impairments were excluded. Data on sociodemographic variables, physical activity, tobacco/alcohol use, BMI, diagnosis of diabetes, hypertension, and other medical comorbidities were collected. Lipid profiles were measured from fasting peripheral venous blood samples using standard laboratory techniques and lipid abnormalities were classified based on the NCEP ATP-III criteria. Proportions of lipid abnormalities were compared between the two populations using the two-proportions Z-test, and risk factors associated with dyslipidemia were analyzed using multivariate logistic regression models.

Results

The prevalence of high total cholesterol (TC), and low-density lipoprotein cholesterol (LDL-c) was significantly greater in the urban than rural population (TC: 37.0% vs. 28.4% p < 0.001 and LDL-c: 33.5% vs. 26.8%, p < 0.01, respectively), while the prevalence of low high-density lipoprotein cholesterol (HDL-c: 72.4% vs. 44.2%, p <0.001), high triglycerides (TG: 45.7% vs. 38.6%, p <0.01) and lipid risk ratios (TC/HDL-c, TG/HDL-c and LDL-c/HDL-c) was higher in the rural than urban population. Females in both urban and rural populations were at a higher risk of having multiple lipid abnormalities. For the other risk factors assessed, while diabetes, overweight, obesity, and physical inactivity were associated with increased risks for certain lipid abnormalities, these associations were less pronounced in the urban population.

Conclusions

Aging Indians, both in rural and urban settings, have an alarmingly high prevalence of lipid abnormalities. Considering an elevated cardiovascular disease risk associated with lipid abnormalities, targeted interventions towards these communities are necessary to reduce the disease burden.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-22625-3.

Background

Lipids are a diverse group of organic biomolecules that play a crucial role in the human body and are essential for maintaining normal physiological functioning [1]. Important classes of lipids include fatty acids, phospholipids, cholesterols, etc. Fatty acids are commonly stored as triglycerides and are an important source of energy, whereas cholesterols are essential components of the cell membrane and are necessary for the synthesis of bile acid, steroid hormones, and vitamin D [2].

Dysregulation in lipid metabolism, termed dyslipidemia, is characterized by one or more of the following abnormalities in the plasma lipid parameters, namely high levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c), and triglycerides (TG), and low levels of high-density lipoprotein cholesterol (HDL-c) [3]. The most commonly used set of criteria for diagnosis of dyslipidemia is the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP-III) [4]. Dyslipidemia is implicated in the pathophysiology of several diseases, such as atherosclerotic cardiovascular disease (ASCVD) [5], stroke [6], diabetes [7], renal disease [8], chronic inflammatory conditions [9], osteoporosis [10], amongst several others. In addition to the individual lipid parameters, lipid risk ratios such as TC/HDL-c and TG/HDL-c ratio have been shown to be good predictors of cardiovascular disease [11].

The global burden of dyslipidemia has risen considerably in the recent few decades [12]. As of 2019, elevated LDLc was the ninth leading cause of death worldwide. Further, it was estimated that between 1990 to 2019, there was a 46% increase in the absolute annual burden from deaths and DALYs (disability-adjusted life years) attributed to high LDL-c [13]. However, the prevalence and patterns of dyslipidemia vary significantly across different populations globally [14–16]. Additionally, there is a research gap in the studies on dyslipidemia, with a notable disparity between high-income countries (HICs) and low- and middle-income countries (LMICs). The recent Indian Council of Medical Research-India Diabetes (ICMR-INDIAB-17) study estimates an alarmingly high prevalence of dyslipidemia in India at 81.2% [17] compared to their previous report of a prevalence rate of 79% [18], with a significantly higher prevalence in the urban than in the rural population. Even though epidemiological studies from different regions and sites of India confirm an increasing trend in lipid abnormalities at the population level [19–21], rural-urban differences in the prevalence of dyslipidemia are not so well-established, especially in the older population.

Considerable changes in lipid metabolism are observed during aging. There are changes in the activity of enzymes involved in lipid synthesis and metabolism, alterations in the levels and activity of lipoproteins (which transport lipids in the blood), and a decrease in the capacity of organs to use lipids as energy sources, thereby resulting in lipid accumulation [22]. Overall, these changes in lipid metabolism with aging can lead to a higher risk of metabolic disorders and cardiovascular diseases [23], cerebrovascular diseases [24] and dementia [25].

Nevertheless, these age-related alterations in lipid metabolism are not uniform and are influenced by genetic and several environmental factors. Lifestyle factors such as diet [26], physical activity [27], smoking and alcohol use [28], and psychological stress [29] are crucial in regulating lipid levels. Interestingly, rural and urban residency could have a substantial impact on all the above-mentioned lifestyle factors. Additionally, the interaction among historical, socio-cultural, and economic factors in different parts of the world could further contribute to unique variations in rural and urban lifestyles, leading to population-specific rural–urban differences in the prevalence of dyslipidemia.

India, being a nation of great diversity, also has a stark rural–urban divide. Rural and urban Indians not only differ in terms of the lifestyle factors discussed above but also have disparities in educational attainment, economic status, health-related awareness, and access to modern healthcare facilities. Urban residents have been typically considered to have a fast-paced life, with less healthy food habits and lesser physical activity, potentially making them more prone to dyslipidemia. However, the ongoing rapid urbanization and demographic shifts in India are challenging this notion. Therefore, it is essential to examine the prevalence of dyslipidemia, individual lipid parameter abnormalities, and abnormal lipid risk ratios between rural and urban Indian populations through parallelly conducted studies utilizing robust clinical and biochemical assessments.

The Centre for Brain Research-Srinivaspura Aging Neuro Senescence and COGnition (CBR-SANSCOG) study and the CBR-Tata Longitudinal Study of Aging (CBR-TLSA) are two ongoing, prospective, aging cohort studies conducted in rural and urban populations respectively in southern India. They are primarily designed to identify risk and protective factors associated with dementia and have harmonized study protocols that include multimodal assessments. The current study aims to compare the prevalence of dyslipidemia and abnormal lipid risk ratios in these two distinct populations.

Methods

Study design and participants

This study utilized cross-sectional data from the baseline assessments of CBR-SANSCOG and CBR-TLSA cohorts. Both these studies recruit community-dwelling individuals aged 45 years and above, who are residents of their respective study sites – villages in Srinivaspura and Bangalore city. According to the most recent Census of India (2011), an area is designated as 'urban' if it meets the following criteria: a population of at least 5,000 people, a population density of at least 400 people per square kilometer, and at least 75% of the primary working population engaged in non-agricultural activities. Areas that do not meet these conditions are classified as 'rural' [30]. Based on these definitions, the study locations where participants were recruited – the villages of Srinivaspura and the city of Bangalore – are classified as rural and urban, respectively.

The CBR-SANSCOG cohort participants are a rural community of primarily mango cultivators and are long-term settlers with a very low rate of migration. These rural participants mostly come from a lower socio-economic background and have lower education. On the other hand the CBR-TLSA cohort participants are long-term residents of Bangalore and mostly belong to a middle/high socio-economic background and are highly educated.

For the CBR-SANSCOG study, an area sampling strategy is followed where villages are grouped under Primary Health Centres (PHCs), which include a number of villages, and willing participants are recruited from each PHC cluster under the Srinivaspura 'taluk' (equivalent to a subdistrict). CBR-TLSA participants are recruited following a convenience sampling strategy, where participants are informed of the study through advertisements, social media, and awareness camps, and willing participants are recruited.

Individuals with dementia, severe or terminal medical illnesses, severe psychiatric disorders, and those with severe visual or hearing, or locomotor impairments that would affect study assessments are excluded. Both CBR-SANSCOG and CBR-TLSA participants undergo comprehensive clinical, cognitive, biochemical, genetic, and neuroimaging assessments as part of the parent cohorts and are followed up longitudinally. The protocols for these studies have been described elsewhere [31–33].

For the present study, we included 4913 rural (CBR-SANSCOG cohort) participants who completed their baseline assessments between January 4, 2018 (starting of CBR-SANSCOG cohort) and October 31, 2022 and 1089 urban (CBR-TLSA cohort) participants who completed their baseline assessments between July 1, 2015 (starting of CBR-TLSA cohort) and October 31, 2022. After the removal of missing values, the final sample size of CBR-SANSCOG and TLSA were 2797 and 430, respectively. The numbers and proportions of missing values for the different parameters are reported in the supplementary file (Supplementary Table 1).

Ethical statement

CBR-SANSCOG and CBR-TLSA studies have been cleared by the Institutional Human Ethics Committee of the Centre for Brain Research. Written, informed consent was taken from all the participants prior to participation in the studies. All data were anonymized to ensure the privacy of the participants.

Estimation of lipid parameters

Trained phlebotomists collected 15 mL of peripheral venous blood, with a single prick, from each fasting (overnight) participant. Serum separation was done by centrifugation of blood at 2000 rpm for 10 min. Lipid parameters were examined by standard laboratory techniques at our nationally accredited collaborating laboratories, as described below.

Total cholesterol

Serum cholesterol was measured by colorimetric enzymatic method, wherein cholesterol esterase cleaves cholesterol esters into free cholesterol and fatty acids. Further, cholesterol oxidase catalyses the oxidation of cholesterol forming hydrogen peroxide, which then forms a red quinone-imine dye combining phenol and 4-aminophenazone. The colour intensity of the dye is directly proportional to the cholesterol concentration, which was measured by Cobas autoanalyzer.

HDL-c

In the same autoanalyzer as above, PEG-modified cholesterol esterase and cholesterol oxidase enzymes were used for the direct determination of HDL-cholesterol. These enzymes have selective catalytic activities for lipoprotein fractions with higher preference for HDL. Using this, the hydrogen peroxide generated after the oxidation of cholesterol reacts with 4-amino-antipyrine and HSDA to form purple-blue dye, the colour intensity of which, is directly proportional to the HDL-c concentration.

LDL-c

The automated method for estimation of LDL-cholesterol was selective micellar solubilization by a non-ionic detergent and interaction with a sugar compound. The enzyme reactions to lipoproteins other than LDL are inhibited by the surfactant and the sugar compound. Subsequently, cholesterol esterase and cholesterol oxidase are added sequentially, leading to the generation of hydrogen peroxide. This, in turn, forms purple-blue dye, whose colour intensity is directly proportional to the LDL-c concentration.

Triglycerides

Triglycerides estimation in human serum using Roche Cobas c systems was based on Wahlefeld’s method. This method uses lipoprotein lipase from microorganisms to completely hydrolyse triglycerides to glycerol, followed by oxidation to dihydroxyacetone phosphate and hydrogen peroxide. The hydrogen peroxide produced reacts with 4-aminophenazone and 4-chlorophenol under the catalytic action of peroxide to form red dye (Trinder endpoint reaction). The colour intensity of this dye is directly proportional to the Triglycerides concentration.

Diagnosis of dyslipidemia

Dyslipidemia was defined as the presence of one or more lipid abnormalities as per the NCEP ATP-III criteria and lipid risk ratios were calculated from the individual lipid parameters [4, 34].

Parameters	Cut-off Level (mg/dl)
Low-density lipoprotein cholesterol (LDL-C)	≥ 130
Total cholesterol	≥ 200
Triglycerides (TG)	≥ 150
High-density lipoprotein (HDL-C) cholesterol	< 40/< 50
TC/HDL-c ratio	≥ 4.5
TG/HDL-c ratio	≥ 4.7 (men)/3.7 (women)
LDL-c/HDL-c ratio	≥ 3
nonHDL-c (TC-HDL-c)	≥ 130

Risk factors of dyslipidemia

Multiple risk factors associated with dyslipidemia were considered, namely sociodemographic factors such as place of residence, age, and sex, lifestyle-associated factors such as body mass index (BMI), self-reported physical activity, current tobacco and current alcohol use, and comorbidities (diabetes and hypertension, self-reported and/or objectively measured).

Statistical analyses

Data analyses were performed using SPSS version 29 and R environment version 4.3.2. After the removal of missing values, the final sample sizes of CBR-SANSCOG and CBR-TLSA were 2797 and 430, respectively. Baseline characteristics were established using appropriate tests: Welch’s two-sample t-test for continuous variables and Pearson’s Chi-square test for categorical variables. Lipid parameters when taken as continuous variables were skewed and hence, were presented as medians with interquartile ranges (IQR) and compared using the Mann–Whitney U test. Categorical variables were compared in terms of proportions using Z-tests for proportions. As there was no established cut-off for LDL-c/HDL-c, this was estimated from our population-based risk factors following the method used by Chen et al. 2016. [35] (Supplementary Table 2).

Logistic regression was performed to calculate odds ratios based on each of the predictor variables while adjusting the model for the rest of the variables. The independent predictor variables were further categorized as: place of residence (rural [reference], urban), sex (male [reference], female), age (continuous), BMI (categorized into underweight, normal [reference], overweight and obese), physical activity (active [reference], inactive), tobacco use (non-users [reference], currently using), and alcohol use (non-users [reference], currently using). The outcome variables were dyslipidaemia, and abnormal TC, TG, HDL-c, LDL-c, TC/HDL-c ratio, TG/HDL-c ratio, LDL-c/HDL-c ratio, and non-HDL-c. Results are reported as adjusted odds ratio (OR) with 95% confidence interval [OR (95% CI)]. In the urban population only two individuals were there in the underweight category of BMI and hence were excluded from the further analysis.

Results

Participant characteristics

The mean age of the rural sample was 59 years (SD = 10) and that of the urban sample was 66 years (SD = 7); this difference was statistically significant (p < 0.001). Sex differences between rural and urban populations were not statistically significant. The mean years of education was significantly lower (p < 0.001) in the rural cohort (4.7 years, SD = 4.8) as compared to the urban cohort (15 years, SD = 3.6). The rural cohort had a significantly lower (p < 0.001) proportion (17%) of individuals with household income > 50,000 INR than that in the urban cohort (86%). Further, the rural participants had significantly lower proportions of alcohol use, hypertension, diabetes, obesity, physical inactivity, cardiac illness, and stroke than their urban counterparts (Table 1).

Table 1.

Socio-demographic variables in rural and urban populations

	Rural/SANSCOG n = 2797	Urban/TLSA n = 430	p-values
Age (mean, SD)	59(10)	66(7)	< 0.001
Education (mean, SD)	4.7(4.8)	15(3.6)	< 0.001
Sex (%, n)			0.8
Males	51 (1439)	52 (224)
Females	49 (1358)	48 (206)
Income (%, n)			< 0.001
< 25,000	56 (1567)	3.5 (15)
25,000–50000	27 (742)	11 (46)
> 50,000	17 (488)	86 (369)
Smoking (%, n)			< 0.001
Yes	34.2 (959)	5.5 (24)
Alcohol use (%, n)			< 0.001
Yes	11 (309)	30 (127)
Hypertension (%, n)			< 0.001
Yes	32 (901)	74 (318)
Diabetes			< 0.001
Yes	20 (546)	38 (162)
Obesity (%, n)			< 0.001
Yes	31 (855)	67 (287)
Physical inactivity (%, n)			< 0.001
Yes	3.6 (102)	42 (180)
Cardiac Illness (%, n)			< 0.001
Yes	2.2 (62)	8.4 (36)
Stroke (%, n)			0.9
Yes	0.6 (16)	0.5 (2)

Hypertension: self-reported diagnosis/drug treatment and/or BP ≥ 140 / 90 mm Hg; Diabetes: self-reported diagnosis/drug treatment and/or fasting glucose ≥ 126 mg/dl; Obesity: BMI ≥ 25; Physical Inactivity: < 600 METs/week on the Global Physical Activity Questionnaire; Cardiac illness: self-reported diagnosis; Stroke: self-reported diagnosis. Welch’s two-sample t-test was used to compare continuous variables and Pearson’s Chi-square test was used to compare categorical variables

Comparison of abnormal lipid parameters in rural and urban participants (Table 2)

Table 2.

Comparison of proportions of abnormal levels of lipid parameters between in rural and urban populations

Lipid Parameters	Gender	Rural (n = 2797)	Urban(n = 430)	p -values
Abnormal TC (%)	Males	21.8	28.6	< 0.05
	Females	35.3	46.1	< 0.01
	Overall	28.4	37.0	< 0.001
Abnormal TG (%)	Males	46.8	34.8	< 0.001
	Females	44.6	42.7	0.62
	Overall	45.7	38.6	< 0.01
Abnormal LDL-c (%)	Males	19.9	27.2	< 0.05
	Females	34.2	40.3	0.10
	Overall	26.8	33.5	< 0.01
Abnormal HDL-c (%)	Males	67.4	41.1	< 0.001
	Females	77.6	47.6	< 0.001
	Overall	72.4	44.2	< 0.001
Dyslipidemia (%)	Males	82.8	66.1	< 0.001
	Females	89.5	79.6	< 0.001
	Overall	86.1	72.6	< 0.001
Abnormal TC/HDL-c (%)	Males	55.6	43.8	< 0.001
	Females	49.8	31.1	< 0.001
	Overall	52.8	37.7	< 0.001
Abnormal TG/HDL-c (%)	Males	40.5	23.7	< 0.001
	Females	44.1	29.1	< 0.001
	Overall	42.3	26.3	< 0.001
Abnormal LDL-c/HDL-c (%)	Males	24.5	22.3	0.54
	Females	23.9	14.6	< 0.01
	Overall	24.2	18.6	< 0.05
Abnormal non-HDL-c (%)	Males	53.3	51.3	0.63
	Females	65.3	63.6	0.68
	Overall	59.1	57.2	0.48

TC Total Cholesterol, TG Triglyceride, LDL-c Low density lipoprotein cholesterol, HDL-c High density lipoprotein cholesterol. Proportions were compared using Z-tests for proportions

The median total cholesterol level was significantly higher (p < 0.05) in the urban cohort (185.3 mg/dL) as compared to the rural cohort (178 mg/dL). On the contrary, rural participants had higher median values than urban for HDL (45.8 mg/dL vs. 38.7 mg/dL, p < 0.001) as displayed in Table 3.

Table 3.

Comparison of median values of lipid parameters between rural and urban populations

Variable	Median		IQR		p-values
	Rural/SANSCOG	Urban/TLSA	Rural/SANSCOG	Urban/TLSA
TC	178	185.3	50.5	66.2	< 0.01
TG	141	132.6	98	81.4	> 0.05
LDL-c	109.3	114.5	42	59.1	> 0.05
HDL-c	38.7	45.8	12.8	14.4	< 0.001

IQR Interquartile range, TC Total Cholesterol, TG Triglyceride, LDL-c Low density lipoprotein cholesterol, HDL-c High density lipoprotein cholesterol. Medians with IQR were compared using the Mann–Whitney U test.

Total cholesterol (TC)

Urban participants had a significantly higher proportion of abnormally high TC than rural participants (37% vs. 28.4%, p < 0.001) (Table 2). A similar trend was seen among males (urban: 28.6% vs. rural: 21.8%, p < 0.05) and females (urban: 46.1% vs. rural: 35.3%, p < 0.01) (Fig. 1).

Fig. 1.

Rural–urban comparison of the proportion of lipid abnormalities. TC- Total Cholesterol, TG- Triglycerides, LDL- Low density lipoprotein cholesterol, HDL- High density lipoprotein cholesterol. p-values: * < 0.05, ** < 0.01, *** < 0.001, ns- not significant

Low-density lipoprotein cholesterol (LDL-c)

Urban participants had a significantly higher proportion of abnormally high LDL-c than rural participants (33.5% vs. 26.8%, p < 0.01) (Table 2). A similar trend was seen among males (urban: 27.2% vs. rural: 19.9%, p < 0.05). However, there were no significant differences in the proportions of abnormal LDL-c between rural and urban women (Fig. 1).

High-density lipoprotein cholesterol (HDL-c)

Contrary to the above trends observed for total cholesterol and LDL-c, rural participants had a significantly higher proportion of abnormally low HDL-c than urban participants (72.4% vs. 44.2%, p < 0.001) (Table 2). This difference was statistically significant among males (rural: 67.4% vs. urban: 41.1%, p < 0.001) as well as among females (rural: 77.6% vs. urban: 47.6%, p < 0.001 (Fig. 1).

Triglycerides (TG)

Rural participants had a significantly higher proportion of abnormal TG than urban participants (45.7% vs. 38.6%, p < 0.01) (Table 2). This difference was statistically significant only among males (rural: 46.8% vs. urban: 34.8%, p < 0.001) and not among females (rural: 44.6% vs. urban: 42.7%, p = 0.62) (Fig. 1).

Dyslipidemia

The overall proportion of individuals with dyslipidemia (computed as self-report and/or abnormality in any one of the above four lipid parameters) was significantly higher (< 0.001) in the rural (86.1%) than the urban participants (72.6%) (Table 2). On gender stratification, the same trend was observed among males (rural: 82.8% vs. 66.1%, p < 0.001) and females (89.5% vs. 79.6%, p < 0.001) (Fig. 1).

Lipid risk ratios and non-HDL-c

The proportion of participants with abnormal lipid TC/HDL-c ratio was significantly higher (p < 0.001) in the rural (52.8%) as compared to the urban sample (37.7%) (Table 2). Similarly, the proportion of participants with abnormal lipid TG/HDL-c ratio was significantly higher (p < 0.001) among the rural (42.3%) than the urban participants (26.3%). Similar trend was observed for abnormal LDL-c/HDL-c ratio (p < 0.05) as well (24.2% in the rural compared to 18.6% in the urban). For both the abnormal TC/HDL-c and TG/HDL-c ratios, rural males had a significantly higher proportion than urban males and the same trend was observed when comparing rural and urban females (Fig. 2). However, there was no difference among rural and urban males in abnormal LDL-c/HDL-c ratio, but the trends observed in the females were similar to the other two ratios.

Fig. 2.

Rural–urban comparison of the proportions of lipid risk ratios and nonHDL. TG/HDL: Triglyceride/High-density lipoprotein cholesterol ratio, TC/HDL: Total cholesterol/High-density lipoprotein cholesterol ratio, LDL/HDL: Low-density lipoprotein cholesterol/High-density lipoprotein cholesterol ratio, nonHDL: non- High-density lipoprotein cholesterol [TC- HDLc]. p-values: * < 0.05, ** < 0.01, *** < 0.001, ns- not significant

Risk factors associated with lipid abnormalities

We ran separate multivariate logistic regression models with various risk factors associated with all lipid parameters, dyslipidemia and lipid risk ratios, categorizing them into normal and abnormal groups. Place of residence was included as an interaction term in each of these models to assess whether place of residence (rural or urban) significantly moderated the association between the risk factors and lipid abnormalities, i.e. the risk of having lipid abnormalities predicted by each of these factors varied depending on the place of residence of the participants. We found a significant interaction between place of residence and multiple risk factors (Supplementary Table 3–11 and supplementary Fig. 2–10). Given this significant interaction, we stratified the population based on the place of residence and reported the associations between risk factors and lipid abnormalities as ORs (Odds Ratios) and 95% Confidence intervals (Figs. 3, 4, 5, 6 and 7).

Fig. 3.

Associations between various risk factors and TC across rural and urban poopulations

Fig. 4.

Associations between various risk factors and TG across rural and urban populations

Fig. 5.

Associations between various risk factors and LDL-c across rural and urban populations

Fig. 6.

Associations between various risk factors and HDL-c across rural and urban populations

Fig. 7.

Associations between various risk factors and dyslipidemia across rural and urban populations

Females had a higher risk of having lipid abnormalities in both the populations across multiple parameters such as TC [rural: 2.3 (1.91, 2.78), urban: 2.37 (1.43, 3.99)], LDL-c [rural: 2.34 (1.93, 2.83), urban: 2.14 (1.27, 3.63)], HDL-c [rural 1.58 (1.31, 1.91)], and dyslipidemia [rural: 1.9 (1.49, 2.43), urban: 1.76 (1.04, 3)]. Diabetes was associated with an increased risk of lipid abnormalities in the rural population for TC [1.43 (1.16, 1.77], TG [2.59 (2.1, 3.2)], HDL-c [1.46 (1.15, 1.86)], and dyslipidemia [1.92 (1.37, 2.75)] and a decreased risk of lipid abnormalities in the urban population for TC [0.32, (0.2, 0.51)] and LDL-c [0.36 (0.22, 0.57)]. Being underweight was consistently associated with a lesser risk for lipid abnormalities in the rural population, whereas being overweight and obese were associated with an increased risk of abnormal TG, HDL-c, and dyslipidemia in both populations. The findings for the association between various risk factors and the lipid risk ratios and non-HDL-c have been provided in the Supplementary Figs. 11–14.

Discussion

Our study found significant rural–urban differences in the prevalence of lipid abnormalities in this middle-aged and older Indian population. The high prevalence of abnormal lipid levels in both our rural and urban cohorts is alarming, yet in line with a recent nationwide study among Indian adults (20 + years), namely the ICMR-INDIAB-17 study [17]. Among the individual lipid parameters, our study indicated that the prevalence of abnormal TG and HDL-c was higher in the rural than the urban population, whereas the inverse trend (urban > rural) was observed for the prevalence of abnormal TC and LDL-c.

A higher prevalence of abnormal TC and LDL-c in the urban population has been reported in multiple other studies internationally [36–38] and in a previous study from India [18] as well. Likewise, a higher prevalence of abnormal TG and HDL-c in the rural population has been reported in certain other studies [39, 40]. However, there is some disagreement between our findings and those of other studies regarding certain lipid parameters. In India, findings of the ICMR-INDIAB -17 study observed that the prevalence of all the individual lipid parameters was higher in the urban population than the rural – abnormal TC (27.4% in urban and 22.3% in rural), LDL-c (23.5% in urban and 19.6% in rural), HDL-c (68.1% in urban and 66.3% in rural) and TG (36.4% in urban and 30% in rural) in adults aged 20 years and above [17]. Possible reasons for these contrasting findings could be the varied age groups of the populations studied and differences in diagnostic criteria. Geographical, socio-cultural, and genetic variations could also account for the differences in prevalence patterns since there exists tremendous diversity across India with respect to the above-mentioned factors.

With the rapidly ongoing urbanization in India, there is a noticeable shift in dietary habits, even in rural areas. Increasingly, these communities are gaining access to unhealthy food options. This shift is likely to contribute to an increase in the prevalence of lipid abnormalities in the rural community. The ICMR-INDIAB-17 study [17] revealed that the prevalence of several lipid abnormalities increased substantially as compared to the ICMR-INDIAB-14 [18] study findings around a decade ago (TC: 22.3% from 11.7%, TG: 30% from 26% and LDL-c 19.6% from 10.1%). It is also possible that rural individuals could be using cheaper, less refined cooking oils with high saturated fat content (for example, palm oils)—another contributor of lipid abnormalities. Further, the rural population, which is predominantly engaged in farming could have easy access to dairy products, which are rich in saturated fat [41].

The most prevalent individual lipid abnormality in our study was low-HDL-c both in the urban (44.2%) and rural populations (72.4%), with much higher prevalence in the latter group. Different subclasses of HDL-c categorized based on their size differ in their cardioprotective properties. The largest particle, α−1 HDL-c is the most cardioprotective in nature [42]. Studies have shown that Indians have lower concentrations of larger HDL particles compared to Caucasians and that the HDL particle size was smaller in Asian Indians [43]. Along with a very high prevalence of low-HDL-c in the population, lower HDL-c particle size further worsens HDL-c-associated CVD risks. Screening for HDL-c particle size along with its total level in relation to cardiovascular disease, events and mortality might prove to be more beneficial in dealing with CVD burden.

Abnormal TG was the second most prevalent condition in our study, with a significantly higher prevalence in the rural (45.7%) as compared to the urban (38.6%) cohort. While TG in itself is not considered to be atherogenic, its association with remnant particles marks it as an important biomarker for CVD risk assessment [44]. The combined effect of abdominal adiposity, glucose intolerance, low HDL-c, and high TG have been described as the South Asian phenotype [45], which in turn has been associated with higher CVD risk. Concurrent with this concept, Asian-Indians were found to be 5–7 times more susceptible to developing dysglycemia and dyslipidemia compared to other ethnic groups in the USA.

Due to underlying metabolic links, high TG and low-HDL-c often coexist [46]. This condition, also called as atherogenic dyslipidemia, is a common occurrence in South Asians and is often associated with diabetes, metabolic syndrome, and coronary heart disease [47, 48]. The dual burden in terms of the higher prevalence of these two coexisting conditions, also evident by the higher TG/HDL-c ratio among rural than urban individuals, poses a serious threat of CVD risk to the rural population. Studies have reported that such lipid risk ratios better predict CVD risk than individual lipid parameters. In addition, the rural population had a higher prevalence of individuals with abnormal TC/HDL-c ratio and LDLc/HDLc ratio. The prevalence of abnormal non-HDLc, which is often considered as a better predictor of CVDs compared to TC and LDLc, as it gives an estimation of APOB-carrying particles in the blood, was not significantly different across the urban and rural populations.

Our study observed significant rural–urban differences with respect to the socio-demographic variables included in our models. Surprisingly, even though the proportions of well-known risk factors of dyslipidemia, such as physical inactivity, alcohol use, hypertension, diabetes, obesity, and cardiac illness, were higher in the urban population, the proportion of certain lipid abnormalities was higher in the rural population, possibly indicating a role of other factors such as genetic (polymorphisms) and other lifestyle-associated factors.

On exploring the potential risk factors associated with lipid abnormalities, we found that females had a higher risk of lipid abnormalities across the rural (TC, LDL-c, HDL-c, dyslipidemia, TG/HDL-c ratio, and non-HDL-c) and urban (TC, LDL-c, and dyslipidemia) populations. Sex hormones are important regulators of the plasma lipid profile along with the sex chromosome [49]. Despite higher plasma lipoprotein levels, premenopausal women exhibit a lower incidence of cardiovascular diseases compared to men of the same age, with an approximate 10-year delay in the onset of the first cardiac event [50–52]. However, this risk escalates rapidly following menopause [53]. CVDs are still the leading cause of mortality in women, and dyslipidemia contributes the highest population-adjusted risk for CVDs compared with all other risk factors of atherosclerotic CVD [54, 55]. Therefore, identifying at-risk females both at pre- and post-menopausal stages becomes essential for CVD risk management.

Further, in the rural population, diabetes was associated with an increased risk of lipid abnormalities, including elevated TC, TG, dyslipidemia, lipid risk ratios, and non-HDL-c. In contrast, in the urban population, diabetes was associated with a decreased risk of certain lipid abnormalities such as TC, LDL-c, the LDL-c/HDL-c ratio, and non-HDL-c. Insulin resistance (IR) is known to affect lipid and lipoprotein metabolism and thus, plays a crucial role in the development of dyslipidemia [56]. Additionally, the accumulation of lipids in peripheral tissues induces IR [57]. Consequently, the underlying pathophysiological mechanisms of IR and lipid accumulation operate in tandem and, without proper management, may lead to adverse outcomes. These contrasting findings between rural and urban populations could be explained by better diabetes management in urban areas. Increased awareness of the condition and the use of antidiabetic medication can not only improve glycemic control but also influence plasma lipoprotein levels, thereby reducing the risk of certain lipid abnormalities.

The strength of our study lies in its community-based design and large sample sizes from the rural and urban Indian populations. Further, harmonized assessment protocols between the rural and urban cohorts and robust lipid profile assessments from fasting blood samples (direct measurement of LDL-c instead of a derived measure from other lipid parameters) are added strengths. Moreover, the rural study site in Srinivaspura is one of the major mango cultivation hubs in India, and the resident population has settled here for several generations. So, the migration rates are low, and this is a fairly homogenous rural Indian population.

However, our study has certain limitations as well. Firstly, caution needs to be exercised regarding the generalizability of the findings, particularly concerning our urban population, which was recruited through convenience sampling. Further, the rural and urban sample sizes were disproportionate, and the limited sample size of the urban population may not have been sufficient to identify the different risk factors associated with lipid abnormalities. Additionally, we did not conduct a detailed dietary assessment, which could have potentially shed light on the reason behind the rural–urban differences observed between these two populations. Also, levels of apolipoproteins such as ApoB, Apo(a) were not measured, which could give additional information regarding the dysregulation of lipid metabolism in the population.

Cardiovascular disease and associated mortality are on the rise in the urban as well as the rural population in India. Against this backdrop, our study not only points to the high prevalence of lipid abnormalities in both these populations but also brings to light the differential prevalence patterns of abnormal lipid parameters across these two diverse populations. We call for urgent public health measures such as improving awareness, initiating community screening programs, and implementing intervention strategies that include both pharmacological and lifestyle-based management.

Rural Indians face challenges such as poor health awareness, financial constraints, and limited access to treatment, therefore the undiagnosed/underdiagnosed lipid abnormalities potentially in combination with other comorbidities can lead to adverse cardiac events. Contrarily, urban individuals must tackle the higher prevalence of comorbidities such as diabetes, hypertension, and obesity, along with the higher proportions of TC and LDL-c (the most well-studied CVD-associated lipid abnormality), which potentially puts them at higher CVD risk. Long-term, prospective studies are required in both rural and urban Indian populations to evaluate the extent of the impact of dyslipidemia on multiple health dimensions, quality of life, and socio-occupational functioning, as age advances. CVD risk assessment tools for Indians must also be revisited considering the high prevalence of dyslipidemia and the benefits of using lipid-lowering drugs in this population. A recent survey among Indian doctors revealed considerable variability in the management strategies for lipid abnormalities, especially in estimating cardiovascular risk, targeting non-HDL cholesterol abnormalities, and the use of non-statin treatments [58]. National guidelines for the screening and management of lipid abnormalities curated for rural and urban Indian populations will provide a strong impetus to ensure timely diagnosis and effective management of dyslipidemia in the country.

Conclusion

Pronounced differences in dyslipidemia prevalence and associated factors exist between rural and urban populations in India. Interestingly, despite the presence of a comparatively lesser proportion of conventional risk factors like diabetes, hypertension, obesity, and physical inactivity, the proportions of multiple lipid abnormalities were higher in the rural compared to the urban population. On the other hand, the higher prevalence of LDL-c and other CVD risk factors in the urban population raises significant concern. These results indicate a nuanced interaction of genetic, lifestyle, and dietary influences that vary by place of residence, highlighting the pressing need for public health initiatives to be tailored to the distinct needs of rural and urban populations. Such targeted interventions are essential for reducing the impact of dyslipidemia and its associated cardiovascular disease risks, thereby enhancing health outcomes across India’s diverse aging populations.

Abbreviations

TC

Total Cholesterol

LDL-c

Low-density lipoprotein cholesterol

HDL-c

High-density Lipoprotein cholesterol

TG

Triglyceride

CVD

Cardiovascular disease

ASCVD

Atherosclerotic cardiovascular disease risk

CBR-SANSCOG

Centre for Brain Research-Srinivaspura Aging Neuro Senescence and COGnition study

CBR-TLSA

Centre for Brain Research-Tata Longitudinal Study of Aging

Authors’ contributions

PC, SA, LD, and JSS wrote the main manuscript. SA and PC analyzed the data. PR, JSS, and TGI helped with data curation. All authors reviewed the manuscript.

Funding

The CBR-SANSCOG study is funded by Pratiksha Trust through the Centre for Brain Research. CBR-TLSA is funded by Tata Trusts.

Data availability

The data that support the findings of this study are available from the corresponding author, upon reasonable request, in accordance with the data sharing policies of the Centre for Brain Research and the statutory regulations of the Government of India.

Declarations

Ethics approval and consent to participate

This study involved human participants and was approved by CBR Institutional Ethics Committee. IEC Number: CBR/42/IEC/2022–23. Participants gave informed consent to participate in the study.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.