Association between cardiometabolic phenotypes with atherogenic index of plasma: a cross-sectional study from the Azar cohort population

Abstract

Background

Cardiometabolic phenotypes combine metabolic health and obesity measures for a more accurate cardiovascular risk assessment than body mass index (BMI) alone. The Atherogenic Index of Plasma (AIP), based on triglycerides and high-density lipoprotein cholesterol (HDL-C), is a promising cardiovascular disease (CVD) risk marker. This study explores the relationship between cardiometabolic phenotypes and AIP to improve understanding of their combined predictive value for CVD risk.

Methods

This cross-sectional study analyzed data from 9,515 participants aged 35–55 in the Azar Cohort Study. Metabolic syndrome (MetS) was defined using ATP III criteria. Participants were classified into four phenotypes: metabolically healthy normal weight (MHNW, BMI < 25 kg/m²), metabolically unhealthy normal weight (MUHNW, BMI < 25 kg/m²), metabolically healthy obese (MHO, BMI ≥ 25 kg/m²), and metabolically unhealthy obese (MUHO, BMI ≥ 25 kg/m²). AIP was calculated as the logarithm of triglycerides to HDL. Multinomial regression was used to analyze the relationship between AIP tertiles and phenotypes using both unadjusted and adjusted models. Confounders were controlled for across the three groups.

Results

Among participants, 50.4% were classified as MHO and 28.2% as MUHO. High-risk AIP levels (> 0.21) were found in 79.6% of MUHNW and 64.6% of MUHO, compared to 18.6% of MHO and 13.5% of MHNW. After adjusting for age, physical activity (METs), the wealth score index (WSI), and smoking status, the odds ratios (ORs) for cardiometabolic phenotypes in the high-risk AIP group remained significant. At the high-risk level, MUHNW (OR = 63.49, p < 0.001), MHO (OR = 1.96, p < 0.001), and MUHO (OR = 32.15, p < 0.001) showed significant associations.

Conclusion

In this study, significant associations were found between cardiometabolic phenotypes and AIP. Findings emphasize integrating metabolic health assessments with AIP to improve CVD risk identification. Prospective longitudinal studies are also needed to confirm these associations.

Introduction

Weight assessment is essential for evaluating health and predicting cardiovascular disease (CVD) risk, with body mass index (BMI) calculated as weight divided by height squared being the standard measure [1]. Although BMI categorizes individuals into underweight, normal weight, overweight, and obese groups, it does not capture metabolic health status since it cannot differentiate fat from muscle mass [2]. As a result, metabolic phenotypes combining metabolic syndrome indicators and BMI have been proposed, including metabolically healthy normal weight (MHNW), metabolically unhealthy normal weight (MUHNW), metabolically healthy obese (MHO), and metabolically unhealthy obese (MUHO) [3, 4]. These phenotypes offer a nuanced understanding of CVD-related metabolic health risks, emphasizing the importance of metabolic assessments alongside BMI.

Metabolic health is shaped by a complex interplay of genetic, dietary, and lifestyle factors, making identifying individuals at early risk for chronic diseases challenging [5]. This intricate interaction complicates the identification and investigation of specific causes of metabolic dysfunction, particularly in normal-weight individuals or in understanding the transition from a metabolically healthy to an unhealthy state in obese individuals. These challenges often necessitate longitudinal research and comprehensive physiological assessments [6]. Dyslipidemia, a key feature of metabolic dysfunction, involves elevated levels of triglycerides (TG), very low-density lipoproteins (VLDL), and the smallest subclass of low-density lipoprotein cholesterol (LDL-C) , known as small dense LDL-C (sdLDL-C), along with reduced high-density lipoprotein cholesterol (HDL-C) [7]. SdLDL-C particles are especially atherogenic due to their vulnerability to oxidative modification, which accelerates plaque formation and atherosclerosis. These particles also relate to the fractional esterification rate of HDL-C (FERHDL), reflecting intricate lipid metabolism dynamics [8].

Because sdLDL-C measurement requires costly and complex techniques, such as ultracentrifugation or gradient gel electrophoresis [9], the atherogenic index of plasma (AIP), the base-10 logarithm of the TG-to-HDL-C ratio, serves as a practical surrogate. AIP inversely correlates with sdLDL-C levels and directly associates with lipoprotein particle size, FERHDL, and residual lipoproteinemia, making it a robust indicator of atherogenic dyslipidemia [10, 11].

Importantly, AIP is positively linked to insulin resistance (e.g., Homeostatic Model Assessment of Insulin Resistance [HOMA-IR]),systemic inflammation (e.g., high sensitivity C-reactive protein [hs-CRP]), hypertension, and metabolic syndrome, underscoring its relevance to metabolic health pathways [12–15]. Even among lean adolescents with low baseline risk (AIP < 0.11), elevated AIP predicts future metabolic syndrome development [16]. Since these lipid abnormalities and their metabolic consequences often occur irrespective of BMI, AIP offers a mechanistically grounded marker to differentiate cardiometabolic phenotypes. Its reflection of lipid alterations, insulin resistance, and inflammation supports its utility in identifying metabolically unhealthy obesity and metabolically unhealthy normal weight individuals.Although previous studies, such as the Hoveyzeh Cohort Study in Iran [17], have examined the relationship between AIP and metabolic obesity phenotypes, they primarily treated AIP as a continuous variable without referencing standardized cardiovascular risk cut-offs, like those proposed by Dobiášová [18]. Moreover, despite growing recognition of cardiometabolic phenotypes, limited research has stratified AIP across these groups using clinically meaningful thresholds. This approach enables a more nuanced understanding of how BMI and metabolic status jointly influence AIP levels, moving beyond the limitations of binary classifications based solely on BMI, metabolic syndrome status, or inconsistent cut-offs. By adopting this multidimensional framework, our study addresses a significant gap in the literature by examining the relationship between cardiometabolic phenotypes and the atherogenic index of plasma in the large Azar cohort, thereby contributing to a more comprehensive understanding of cardiovascular risk across metabolic profiles.

Methods and materials

Research design and participants

This cross-sectional study is based on the Azar cohort study, part of the Prospective Epidemiological Research Studies in Iran (PERSIAN) [19]. The comprehensive cohort profile describes the initiation of the Azar cohort in 2014, which consisted of three phases a pilot phase, an enrolment phase, and long-term follow-up of participants [20]. In the present study, data from the pilot phase and enrolment have been used. The conditions for entering the Azar cohort study were: age 35 to 70 years, at least 9 months of permanent residence in Shabestar city, and having at least one Azeri parent. The exclusion criteria included individuals with mental or physical disabilities. All participants provided written informed consent and received detailed information about the study procedure. In the current study, the inclusion criteria were individuals aged 35 to 55 years. Exclusion criteria also included pregnant females, individuals with a history of cancer, those with a total daily energy intake exceeding 4,200 kcal or below 1,200 kcal for males and exceeding 3,500 kcal or below 1,000 kcal for females, as well as those with missing data. Based on these criteria, out of 15,001 participants in the Azar cohort study, a total of 9,515 individuals were included (Fig. 1).

Fig. 1.

Flow chart of participant selection

Ethical approval

This study was approved by the Ethics Committee of the Research Council of Tabriz University of Medical Sciences (No.IR.TBZMED.REC.1402.428).

Data collection

Demographic information, including age, gender, education level, socio-economic status (SES), physical activity, and smoking status, was collected. Education level was classified into four groups based on the level and duration of education. SES classification was performed using the Wealth Score Index (WSI), which includes factors such as housing conditions (e.g., number of rooms, type of homeownership), available infrastructure (drinking water supply, sanitation facilities), car value, household appliances, electronics, and educational resources. WSI was analyzed using multiple correspondence analysis (MCA), allowing WSI to be categorized into five groups, ranging from the poorest to the richest.

Levels of physical activity were evaluated using the metabolic equivalent of task (MET), which measures energy expenditure relative to an individual’s body weight. Specifically, 1 MET corresponds to an oxygen consumption rate of 3.5 mL per kilogram of body weight per minute while at rest [21, 22]. Smoking status was assessed in two categories: non-smoker and smoker. The smoking scale consisted of two options: if respondents reported that they had never smoked, had smoked fewer than 100 cigarettes in their lifetime, or had quit smoking more than a year ago, they were defined as “non-smokers”. If they reported that they were currently smoking or how many cigarettes a day they smoked, they were defined as “smokers” [23].

Biochemical and clinical measurement

Blood samples were obtained following a 12-h fasting period to assess biochemical indicators. TG, total cholesterol (TC), and HDL-C levels were evaluated utilizing enzymatic colorimetric methods, employing commercially available kits from Pars Azmoun in Tehran, Iran. Fasting blood glucose (FBG) was determined through the enzymatic colorimetric method using glucose oxidase. LDL-C was calculated using the traditional Friedewald’s formula.

Blood pressure measurements were taken using a Richter sphygmomanometer by trained personnel, with participants seated and resting for 10 min. Two readings were recorded from each arm, with a one-minute pause between measurements. The average of the second readings from both the right and left arms was calculated to establish the blood pressure level.

Anthropometric measurements

The anthropometric parameters evaluated included height, weight, BMI, waist circumference (WC), and hip circumference (HC). Weight was measured using a Seca digital scale with 0.1 kg accuracy.

Height was measured using a stadiometer with an accuracy of 0.1 cm. WC was measured at the midpoint between the lower costal edge and the iliac crest using a measuring tape. HC was measured around the widest part of the buttocks, ensuring the tape was parallel to the floor. BMI was calculated by dividing weight (kg) by the square of height (m²).

Definition of cardiometabolic phenotypes

Metabolic syndrome (MetS) was defined according to the criteria specified in the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) guidelines [24]. Metabolic syndrome was considered present if at least three of the following.

• FBG ≥ 100 mg/dl or the use of blood glucose-lowering medications.

• Serum TG ≥ 150 mg/dl or the use of TG-lowering medications.

• HDL-C ≤ 40 mg/dl in males and ≤ 50 mg/dl in females, or the use of HDL-C boosting medications.

• Systolic Blood Pressure (SBP) ≥ 130 mmHg or Diastolic Blood Pressure (DBP) ≥ 85 mmHg, or the use of blood pressure-lowering medications.

• Waist circumference >102 cm in males and > 88 in females.

Metabolic phenotypes were defined based on BMI and metabolic status according to the ATP III guidelines [25], as seen below:

1. Metabolically healthy normal weight (MHNW): BMI < 25 kg/m² and fewer than three metabolic syndrome components

2. Metabolically unhealthy normal weight (MUHNW): BMI < 25 kg/m² and at least three metabolic syndrome components

3. Metabolically healthy obese (MHO): BMI ≥ 25 kg/m² and fewer than three metabolic syndrome components

4. Metabolically unhealthy obese (MUHO): BMI ≥ 25 kg/m² and at least three metabolic syndrome components.

Calculation Atherogenic Index of Plasma (AIP)

AIP is calculated as log₁₀ (TG/HDL-C), with TG and HDL-C in mmol/L.

The formula is:

In this study, based on epidemiological data, three risk categories for AIP have been suggested: low risk (AIP < 0.11), intermediate risk (AIP = 0.11–0.21), and high risk (AIP > 0.21).

These thresholds have been associated with cardiovascular risk and characteristics in earlier clinical and epidemiological studies [26, 27].

Statistical analysis

Data analysis was conducted using SPSS version 23.0 software (SPSS Inc., Chicago, IL, USA). The normality of the data was assessed through skewness and kurtosis tests, along with descriptive statistics. Quantitative variables that followed a normal distribution were reported as means ± standard deviations (SD), while those with a non-normal distribution were presented as medians and interquartile ranges (IQR). Qualitative variables were reported as frequencies (percentages).

To compare quantitative variables between groups, one-way analysis of variance (ANOVA) was used, with the least significant difference (LSD) post hoc test applied for multiple comparisons. The Chi-square test and Kruskal–Wallis test was used to compare qualitative variables. Multinomial regression was performed to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between cardiometabolic phenotypes and AIP risk categories (low risk: AIP < 0.11; moderate risk: AIP 0.11–0.21; high risk: AIP > 0.21). This analysis adjusted for potential confounders (age, METs, WSI, and smoking status) across two models (unadjusted and adjusted), with groups categorized by cardiometabolic phenotypes and gender (male and female). The reference categories were MHNW and low-risk AIP. P-values < 0.05 were considered statistically significant.

Results

As shown in Table 1, the largest group was the MHO phenotype, representing 50.4% of the cohort, followed by MUHO at 28.2%, MHNW at 20.3%, and MUHNW at 1.1%. Gender distribution was notable, with males comprising 57.8% of MHNW but only 31.3% of MUHO, while females accounted for 42.2% of MHNW and 68.7% of MUHO. Educational attainment also varied significantly, with the MHNW phenotype having higher rates of diploma and college education (43.6% and 13.7%, respectively) compared to the MUHO phenotype (34% and 7.8%, respectively). Physical activity levels were highest in MHNW and MHO (42.3% and 35%, respectively, reporting high activity) and lowest in MUHO (27.2%). Baseline variables were statistically significant across cardiometabolic phenotypes (p < 0.001). The AIP risk levels distribution further highlighted differences among the cardiometabolic phenotypes. The MUHNW phenotype had the largest proportion classified as high risk (79.6%). In contrast, the MHNW phenotype had the highest percentage of low-risk classification (73.5%). MHO phenotype exhibited 65.3% classified as low risk and 18.6% as high risk. The distribution of moderate-risk classifications varied modestly across cardiometabolic phenotypes, ranging from 9.7% in MUHNW to 17.2% in MUHO (p < 0.001).

Table 1.

Baseline characteristics stratified by cardiometabolic phenotypes in the Azar cohort population

Variables		Total n:9515	MHNW n: 1931(20.3)	MUNHW n: 103 (1.1)	MHO n: 4801(50.4)	MUHO n: 2680 (28.2)	P
		N (%)	N (%)	N (%)	N (%)	N (%)
Gender	Males	3970(41.7)	1116(57.8)	47(45.6)	1969(41.00)	838(31.3)	< 0.001*
	Females	5545(58.3)	815(42.2)	56(54.4)	2832(59.00)	1842(68.7)
Education levels	Illiterate	772(8.1)	111(5.8)	8(7.8)	338(7.00)	315(11.8)	< 0.001**
	Elementary	4018(42.3)	712(36.9)	40(38.8)	2023(42.2)	1243(46.4)
	Diploma	3708(39.0)	841(43.6)	42(40.8)	1913 (39.9)	912(34.00)
	college	1011(10.6)	265(13.7)	13(12.6)	524 (10.9)	209(7.8)
Socio-economic status (WSI)	The lowest	1919(20.2)	445(23.0)	19(18.4)	878(18.3)	577(21.5)	< 0.001**
	Low	1346(14.1)	279(14.5)	6(5.8)	669(14.00)	392(14.6)
	Middle	1895(19.9)	350(18.1)	29(28.2)	1009(21.00)	507(18.9)
	High	2360(24.8)	433(22.4)	27(26.2)	1255(26.1)	645(24.1)
	The highest	1995(21.0)	424(22.0)	22(21.4)	990(20.6)	559(20.9)
Physical activity (METs)	Low	2905(30.5)	526(27.2)	34(33.00)	1400(29.2)	945(35.3)	< 0.001**
	Moderate	3350(35.2)	589(30.5)	36(35.00)	1718(35.8)	1007(37.5)
	High	3260(34.3)	816(42.3)	33(32.00)	1683(35.00)	728(27.2)
Smoking status	No-Smoker	8074(84.9)	1455(75.3)	81(78.6)	4155(86.5)	2383(88.9)	< 0.001*
	Smoker	1441(15.1)	476(24.7)	22(21.4)	646(13.5)	297(11.1)
AIP	low risk	5051(53.1)	4919(73.5)	11(10.7)	3134(65.3)	487(18.2)	< 0.001**
	moderate risk	1498(15.7)	252(13.00)	10(9.7)	775(16.1)	461(17.2)
	high risk	2966(31.2)	260(13.5)	82(79.6)	892(18.6)	1732(64.6)

Statistical significance was determined using the chi-square test, and significant p-values are indicated by an asterisk (*)

Abbreviations: MHNW Metabolically Healthy Normal Weight, MUHN Metabolically Unhealthy Normal Weight, MHO Metabolically Healthy Obese, MUHO Metabolically Unhealthy Obese, WSI wealth score index, METs Metabolic equivalent of task, AIP Atherogenic index of plasma

^**p Kruskal–Wallis test

Figure 2 presents the mean AIP across cardiometabolic phenotypes. The MHNW phenotype showed the lowest mean AIP at 0.004, while the MUHNW phenotype had the highest mean AIP of 0.35. The MUHO phenotype had a slightly lower AIP value of 0.30 compared to the MUHNW phenotype, whereas the MHO phenotype demonstrated a mean AIP of 0.054. This value is higher than that of the MHNW phenotype but lower than that of the MUHNW and MUHO phenotypes. Significant differences among the cardiometabolic phenotypes were observed (P < 0.001).

Fig. 2.

Comparison of the mean AIP among cardiometabolic phenotypes. (*) Indicates a significant difference between the two groups as determined by LSD post hoc analysis one-way ANOVA (p < 0.001)

Table 2 illustrates the baseline clinical characteristics and biochemical data for the cardiometabolic phenotypes. The mean age of cardiometabolic patients was 45 years. The MUHO phenotype was generally older, with a mean age of 46.59 years, compared to 44.05 years in the MHNW phenotype. The MUHO phenotype exhibited the highest values for weight (83.43 ± 12.76 kg), BMI (32.07 ± 4.18), and WC (100.78 ± 8.93 cm), while the MHNW phenotype had the lowest values for these parameters. Blood pressure, FBG, and lipid profiles showed an increasing trend from MHNW to MUHNW and MUHO. For example, the MUHNW and MUHO phenotypes had the highest mean TG levels at 205.40 ± 75.62 and 203.65 ± 106.69 mg/dL, respectively, whereas the MHNW phenotype had the lowest values. HDL-C levels were significantly higher in the MHNW and MHO phenotypes (48.19 ± 10.97 mg/dL and 47.78 ± 10.91 mg/dL, respectively) compared to the MUHNW and MUHO phenotypes, which had the lowest HDL-C levels (38.30 ± 8.09 and 40.99 ± 8.90 mg/dL, respectively). These differences in clinical and biochemical characteristics among the cardiometabolic phenotypes were statistically significant (p < 0.001).

Table 2.

Comparison of anthropometric and biochemical parameters among the studied groups in the Azar cohort population

Cardiometabolic		Total n:9515	MHNW n:1931	MUNHW n:103	MHO n:4801	MUHO n:2680	p
Variables		Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
Age (years)		45.00 ± 5.89	44.05 ± 6.09a	47.49 ± 5.31b,a,c	44.44 ± 5.74c	46.59 ± 5.68d,a,c	< 0.001*
Weight(kg)		76.28 ± 13.55	62.13 ± 8.50a,c,d	63.28 ± 7.99b,c,d	78.25 ± 11.24c,a,d	83.43 ± 12.76d,a,c	< 0.001*
Height(cm)		162.81 ± 9.32	165.82 ± 9.46a	163.50 ± 8.75b,a,d	162.44 ± 9.18c,a,d	161.27 ± 8.99d	< 0.001*
BMI (kg/m2)		28.82 ± 4.89	22.53 ± 1.8	23.59 ± 1.3	29.65 ± 3.6	32.07 ± 4.1	< 0.001*
WC (cm)		93.09 ± 11.08	80.21 ± 7.11a	86.62 ± 6.39b	94.10 ± 8.71c	100.78 ± 8.93d	< 0.001*
HC (cm)		104.96 ± 8.81	95.74 ± 5.01a,c,d	96.19 ± 4.17b,c,d	106.46 ± 7.34c	109.26 ± 8.60d	< 0.001*
SBP (mmHg)		110.99 ± 15.78	104.32 ± 13.55a	117.27 ± 18.44b,a,c	108.83 ± 13.33c	119.43 ± 17.55d,a,c	< 0.001*
DBP (mmHg)		72.98 ± 9.67	68.56 ± 8.60a	75.83 ± 10.94b	72.04 ± 8.66c	77.75 ± 10.07d	< 0.001*
FBG (mg/dl)	Mean ± SD	95.98 ± 29.22	88.08 ± 17.30	122.25 ± 53.54	90.10 ± 18.80	111.20 ± 41.36	< 0.001**
	Median (IQR)	89 (17)	85.50 (14)	103.00 (31)	88.00 (13)	100.00 (27)
TC (mg/dl)		190.91 ± 38.80	180.38 ± 34.91a	197.73 ± 37.00b,a	190.33 ± 36.89c,a,d	199.27 ± 42.72d,a,c	< 0.001*
TG (mg/dl)	Mean ± SD	149.23 ± 84.09	117.46 ± 54.85	205.40 ± 75.62	130.46 ± 62.11	203.65 ± 106.69	< 0.001**
	Median (IQR)	123 (74)	104.00 (36)	196.00 (94)	112.00 (42)	180.00 (97)
LDL-C (mg/dl)		114.79 ± 34.00	108.27 ± 31.56a	117.78 ± 33.80b,a	116.00 ± 33.00c,a	117.20 ± 36.79d,a	< 0.001*
HDL-C(mg/dl)		45.85 ± 10.97	48.19 ± 10.97a, b,d	38.30 ± 8.09b	47.78 ± 10.91c,a,d	40.99 ± 8.90d	< 0.001*

Abbreviations: SD standard deviation, MHNW Metabolically Healthy Normal Weight, MUHNW Metabolically Unhealthy Normal Weight, MHO Metabolically Healthy Obese, MUHO Metabolically Unhealthy Obese, WC waist circumference, HC hip circumference, SBP systolic blood pressure, DBP diastolic blood pressure, FBG fasting blood glucose, TC Total cholesterol, TG triglyceride, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol

^*P-value of the comparison between groups

^*p One Way ANOVA. ^**p Kruskal–Wallis

^aSignificant difference between MHNW, MUHNW, MHO and MUHO

^bSignificant difference between MUHNW, MHNW, MHO and MUHO

^cSignificant difference between MHO, MHNW, MUHNW and MUHO

^dSignificant difference between MUHO, MHNW, MUHNW and MHO

^bacSignificant difference between MUHNW, MHNW and MHO

^dacSignificant difference between MUHO, MHNW and MHO

^acdSignificant difference between MHNW, MHO and MUHNW

^acSignificant difference between MHNW and MHO

^daSignificant difference between MUHNW and MHNW

The results presented in Table 3 indicate that individuals classified as high-risk based on AIP had significantly greater odds of exhibiting the MUHNW and MUHO phenotypes across all analyses. In both unadjusted and adjusted models, high AIP levels were most strongly associated with the MUHNW phenotype, followed by MUHO. Adjustments for age, METs, WSI, and smoking status further strengthened these associations. In gender-specific analyses, notable differences emerged: among males, high AIP was associated with markedly increased odds of MUHNW (adjusted OR = 101.26, p < 0.001) and MUHO (adjusted OR = 46.43, p < 0.001). Among females, the odds were also elevated for MUHNW (adjusted OR = 63.17, p < 0.001) and MUHO (adjusted OR = 35.69, p < 0.001), although the magnitude of association was lower than in males.

Table 3.

The association between cardiometabolic phenotypes and AIP in the Azar cohort population

	AIP Risk levels	MUHNW		MHO		MUHO
		OR (95%CI)	P	OR (95%CI)	P	OR (95%CI)	P
Cardiometabolic phenotypes (9515)	Low Risk	Reference		Reference		Reference
Model 1	Intermediate Risk	5.12 (2.15–12.18)	< 0.001	1.39(1.19–1.62)	< 0.001	5.33(4.43–6.41)	< 0.001
	High Risk	40.68(21.38–77.39)	< 0.001	1.55(1.33–1.80)	< 0.001	19.41(16.44–22.91)	< 0.001
Model 2	Low Risk	Reference		Reference		Reference
	Intermediate Risk	6.04 (2.53–14.42)	< 0.001	1.52(1.30–1.79)	< 0.001	6.40(5.27–7.77)	< 0.001
	High Risk	63.49(32.89–122.56)	< 0.001	1.96(1.67–2.31)	< 0.001	32.15(26.73–38.66)	< 0.001
Male (3970)	Low Risk	Reference		Reference		Reference
Model 1	Intermediate Risk	4.23(0.59–30.22)	0.15	1.80(1.46–2.20)	< 0.001	5.23(3.61–7.58)	< 0.001
	High Risk	79.15(19.01–329.54)	< 0.001	2.86(2.37–3.44)	< 0.001	40.73(30.13–55.05)	< 0.001
Model 2	Low Risk	Reference		Reference		Reference
	Intermediate Risk	4.74(0.66–34.00)	0.12	1.84(1.50–2.27)	< 0.001	5.49(3.77–8.00)	< 0.001
	High Risk	101.26(24.10–425.35)	< 0.001	2.98(2.46–3.61)	< 0.001	46.43(34.05–63.33)	< 0.001
Female (5545)	Low Risk	Reference		Reference		Reference
Model 1	Intermediate Risk	7.84(2.93–20.90)	< 0.001	1.55(1.20–2.01)	< 0.001	7.97(6.06–10.46)	< 0.001
	High Risk	49.87(23.04–107.93)	< 0.001	1.01(0.74–1.37)	0.93	28.09(21.04–37.50)	< 0.001
Model 2	Low Risk	Reference		Reference		Reference
	Intermediate Risk	8.40(3.13–22.54)	< 0.001	1.59(1.22–2.06)	< 0.001	8.73(6.58–11.57)	< 0.001
	High Risk	63.17(28.81–138.48)	< 0.001	1.08(0.79–1.47)	0.6	35.69(26.38–48.27)	< 0.001

AIP Atherogenic Index of Plasm, OR Odd ratio, CI confidence interval, MUHNW Metabolically Unhealthy Normal Weight, MHO Metabolically Healthy Obese, MUHO Metabolically Unhealthy Obese

Model 1: Unadjusted

Model 2: Adjusted model for age, METs, WSI, and Smoking status

MHO was associated with intermediate AIP risk, particularly in adjusted models. Overall, stronger associations between high AIP and MHUNW phenotypes were observed in males, while females showed lower odds ratios than males for MUHNW and MUHO, and a non-significant association for MHO.

Discussion

In this large-scale population-based cohort study, we examined the association between cardiometabolic phenotypes and AIP. Various studies have explored metabolic health using differing criteria, and alternative anthropometric indices have been proposed to assess metabolic status across diverse populations. For example, Haiyang Fang et al. introduced the weight-adjusted waist index, which showed a significant association with CVD and total mortality, independent of BMI [28]. Given the fundamental differences in body composition across racial groups, BMI may not consistently reflect obesity-related risks. For instance, a BMI of 26.9 kg/m² in the Chinese population corresponds to the same diabetes risk as a BMI of 30 kg/m² in white populations [29], which may explain the differences in the prevalence of cardiometabolic phenotypes across studies. For example, our results indicated that the MHO phenotype had the highest frequency (50.4%), while the MUHNW phenotype had the lowest frequency (1.1%). In contrast, previous studies have reported a significantly lower prevalence of MHO (ranging from 13.3% to 15.1%) and a markedly higher prevalence of MUHNW (10.5% to 13.3%) [30, 31]. Discrepancies may stem from several factors, including genetic predispositions, age and sex distribution, lifestyle differences, and methodological variations in defining metabolic health and obesity. For instance, some studies define metabolic unhealthiness based on insulin resistance markers, while others use broader criteria incorporating blood pressure and lipid profiles, potentially influencing phenotype classification [32, 33].

The discrepancy between the frequency of cardiometabolic phenotypes in the present study and those in previous research prompted us to examine the prevalence of AIP risk levels across different cardiometabolic phenotypes. To the best of our knowledge, this study is the first to report the frequency of risk levels across cardiometabolic phenotypes, with no other studies available for direct comparison. When analyzing these risk levels, a higher percentage of participants fell into the low-risk category, consistent with studies by Barua et al. in Bangladesh [34] and Wang et al. in China [35]. This contrasts with findings from Bo et al. in Malaysia [36] and Gebreyesus et al. in Ethiopia [37], where the high-risk category was more prevalent.

These observed similarities might reflect shared lifestyle factors, while the differences could stem from geographical, cultural, and dietary variations, along with methodological differences in research. Additionally, prevalence rates may not be directly comparable across studies since different AIP cutoffs are often employed depending on study design and objectives. Both geographic and environmental influences may also play a crucial role in AIP risk levels, suggesting that some populations may differ in their susceptibility to metabolic changes resulting from dietary alterations [38].

This underscores the need for a comprehensive analysis of the determinants of AIP across various populations. Overall, our findings indicate that understanding AIP risk frequencies can help identify at-risk groups and inform the development of targeted interventions for preventing cardiovascular diseases, taking into account demographic and environmental characteristics to create more effective health strategies.

This study demonstrated a strong and consistent association between high AIP levels and the likelihood of exhibiting metabolically unhealthy phenotypes. Among males, the adjusted odds of MUHNW and MUHO in the high AIP group were 101.26 and 46.43, respectively, while in females, the corresponding adjusted odds were 63.17 and 35.69. These findings indicate that high AIP levels are most strongly associated with the MUHNW phenotype, followed by MUHO, particularly among males. However, since AIP itself reflects the ratio of triglycerides to HDL-C, and the phenotypes are defined partly by metabolic markers related to triglycerides and HDL-C, these results should be interpreted with caution.

Although the MHO phenotype also showed an association with AIP, it was primarily linked to intermediate rather than high-risk AIP levels, and this association was weaker and often non-significant in females. These patterns suggest that AIP can differentiate between metabolic phenotypes within both normal-weight and obese individuals and may serve as a more sensitive marker for identifying metabolically unhealthy individuals, especially among those with normal BMI. Our results highlight a stronger relationship between AIP and metabolically unhealthy phenotypes in males compared to females, consistent with prior findings that show sex-specific differences in lipid metabolism and cardiovascular risk. These findings align with previous studies while providing new insights into how cardiometabolic phenotypes interact with AIP risk. For example, in line with our study, Tien et al. [39], Zhang et al. [40], and Chen et al. [41] conducted studies grouping participants based on metabolic syndrome indicators, revealing that AIP significantly correlated with metabolic syndrome. Additionally, Zhang et al. reported sex-specific variations in AIP’s predictive power for MetS, with stronger effects observed in males and females, respectively. In contrast, Chen et al. reported stronger effects in females and males, respectively, which contradicts our findings. Differences in odds ratios or discrepancies between statistically significant male and female outcomes across various groups may be influenced by factors such as diet, lifestyle, and other environmental variables. AIP measurements can vary depending on the population studied and the cut-off points used, which complicates comparisons across studies. Also, the findings of the present study were consistent with Sadeghi et al. [42], who noted that high AIP levels were linked to an increased risk of CVD and mortality, and Hong et al. [43], who identified AIP as an independent risk factor for CVD in middle-aged males. Both AIP and MetS share key components like triglycerides and HDL-C, contributing to their strong associations. However, limited sample sizes, particularly in the MUHNW phenotype, may overestimate odds ratios and affect generalizability. Therefore, while our findings offer valuable insights, they should be interpreted with caution, and further studies with larger and more diverse populations are essential to validate these associations. This research highlights AIP as a key marker for cardiometabolic risk, especially in metabolically unhealthy individuals. The MUHO and MUHNW phenotypes, particularly in males, are at elevated risk and could benefit from targeted interventions. Future research should investigate the mechanisms behind these associations and explore sex-specific risk reduction strategies.

Our findings indicate that abnormalities in triglyceride and HDL-C metabolism play a key role in atherogenic risk, as reflected by the AIP. Recent studies have found that a genetic predisposition to type 2 diabetes is associated with higher triglycerides and lower HDL-C, whereas a CVD predisposition is linked to chronic inflammation in fact, lipid metabolism disorders have been shown to increase the probability of atherosclerosis in patients with specific conditions, which further supports the role of lipid metabolism in cardiovascular risk [44, 45]. Proteomic biomarkers indicate inflammation driven by TNF-α, ApoB, and other mediators, which disrupt lipid metabolism through oxidative stress and foam cell formation. Consequently, this leads to dysfunctional HDL and elevated triglycerides, increasing cardiovascular risk and further promoting inflammation [46]. While these mechanisms contribute to atherogenic risk, recent evidence suggests that gut microbiota may influence the progression of atherosclerosis by modulating immune responses [47]. Additionally, short-chain fatty acids (SCFAs) may enhance lipid metabolism by improving cholesterol efflux and stimulating bile acid synthesis, potentially improving triglyceride and HDL-C levels [48]. Such effects highlight the potential of dietary and therapeutic approaches in managing high AIP, particularly in dyslipidemia populations. For instance, animal studies on Time-Restricted Feeding (TRF) suggest it may lower visceral fat and improve metabolic health, though it could also decrease HDL-C levels. Pharmacological interventions, such as metformin, support lipid homeostasis indirectly by reducing BMI and blood pressure [49, 50]. Additionally, patient education programs focusing on diet and physical activity play a vital role in improving lipid profiles and managing chronic diseases, especially in diabetic patients. However, age and disease severity significantly influence their effectiveness [51]. Tailoring lipid-reduction strategies to individual characteristics, such as metabolic phenotype and age, can optimize outcomes in reducing AIP and improving cardiometabolic health. Our findings underscore the predictive value of AIP across different BMI categories for cardiovascular risk assessment. The AIP index may help identify at-risk individuals, including those misclassified as metabolically healthy. These implications are crucial for preventing and managing cardiometabolic diseases. However, longitudinal research is necessary to establish causality and assess the long-term effects of AIP on cardiovascular outcomes. Future studies should also examine how to incorporate AIP into routine clinical practice for targeted interventions and early detection in high-risk populations.

Strengths

The present study is the first to use a large sample size to identify the prevalence and association between the risk levels of the AIP in metabolically healthy and unhealthy phenotypes.

Limitations

Due to the study’s cross-sectional design, we cannot establish a causal link between cardiometabolic phenotypes and the AIP. Both the AIP and MetS share key elements, including triglycerides and HDL-C, which contribute to their significant correlation. However, the small sample sizes in the MUHNW phenotype may result in an overestimation of odds ratios and influence the generalizability of our results. While these findings are important, they must be interpreted carefully. This study examined individuals aged 35 to 55, which may limit its findings for younger or older populations, as cardiometabolic risks and AIP differ with age. Participants were categorized as metabolically healthy or unhealthy based on BMI and ATP III criteria, but these may miss factors like insulin resistance and inflammation. Future research should include a wider age range, conduct subgroup analyses, and incorporate additional measures for a more thorough assessment of metabolic health and AIP. Since this investigation was conducted in a population in northwest Iran, the applicability of the results to other regions or countries may be limited.

Conclusion

In this cross-sectional study, after adjusting for confounding factors across two models, the associations between cardiometabolic phenotypes and AIP remained significant, except for the weaker and non-significant associations observed in MHO at the high-risk AIP level in both adjusted and unadjusted models for females. Despite being classified as metabolically healthy, the MHO phenotype still demonstrated an association with cardiovascular risk, particularly in males, highlighting the need for personalized dietary interventions to optimize lipid profiles. Both MUHNW and MUHO phenotypes exhibit higher AIP levels, emphasizing the need to incorporate AIP into routine evaluations, even for a metabolically healthy obese phenotype. While dietary changes could potentially lower atherogenic risk, additional research is required to confirm AIP’s utility in risk assessment and to define appropriate cut-off points. Understanding the dynamics of MUHNW and MHO may inform public health strategies, as elevated AIP could indicate an increased risk for cardiovascular disease and a potential transition to the MUHO phenotype. Longitudinal studies are needed to explore AIP’s role in this progression.

Abbreviations

AIP

Atherogenic index of plasma

ABCA1

ATP binding cassette transporter A1

ApoB

Apolipoprotein B

BMI

Body mass index

DBP

Diastolic Blood Pressure

FBG

Fasting blood glucose

FERHDL

Fractional esterification rate of HDL-C

HC

Hip circumference

HDL-C

High-density lipoprotein cholesterol

MCA

Multiple correspondence analysis

METs

Metabolic equivalents

MHNW

Metabolically healthy normal weight

MHO

Metabolically healthy obese

MUHNW

Metabolically unhealthy normal weight

MUHO

Metabolically unhealthy obese

SBP

Systolic Blood Pressure

SCFAs

Short-chain fatty acids

sdLDL-C

Small dense LDL sdLDL-C

FERHDL

.

T2D

Type 2 diabetes

TG

Triglyceride

TNF-α

Tumor necrosis factor-α

TRF

Time-Restricted Feeding

WC

Waist circumference

WSI

Wealth Score Index

Authors’ contributions

S.S. designed the research. RM. and E.F. supervised the project,validated the data, and conducted the formal analysis.E.F.curated the data.S.S. analysed the data and wrote the paper, while R.M. and E.F. reviewed and edited the manuscript. All authors reviewed and approved the final manuscript.

Funding

This study was supported by the liver and gastrointestinal diseases research centre (Grant No. 700/108 on 14 March 2016), Tabriz University of Medical Sciences. The funder had no role in the study design, data analysis, interpretation and writing of the manuscript in this study. The Iranian Ministry of Health and Medical Education has contributed to the funding used in the PERSIAN Cohort through Grant no.700/534”. The funder had no role in the study design, data analysis, interpretation and writing of the manuscript in this study.

Data availability

The data that support the findings of this study are available from [Vice Chancellor for Research] but restrictions apply to the availability of these data,which were used under license for the current study, and so are not publicly available.Data are however available from the authors upon reasonable request and with permission of [Vice Chancellor for Research].

Declarations

Ethics approval and consent to participate

This study was approved by the Ethical Board of the Research Council of Tabriz University of Medical Sciences (No.IR.TBZMED.REC.1402.428).

Consent for publication

At the beginning of the cohort study, written informed consent for publication was obtained from all participants or their legal guardians.

Competing interests

The authors declare no competing interests.