Identification of metabolic syndrome using lipid accumulation product and cardiometabolic index based on NHANES data from 2005 to 2018

Abstract

Background

Numerous studies indicate that visceral adipose tissue (VAT) significantly contribute to metabolic syndrome (MetS) development. This study aims to assess the distinguishing value of novel obesity markers, specifically lipid accumulation products (LAP) and cardiometabolic index (CMI), in relation to MetS. Considering the gender disparity in MetS prevalence, it is essential to explore whether LAP and CMI exhibit differential distinguishing capabilities by gender.

Method

The investigation included a total of 11,687 qualified individuals who participated in the NHANES survey spanning a 14-year period from 2005 to 2018. Biochemical analysis of blood and body measurements were utilized to determine LAP and CMI values for each participant. Inclusion of gender as a variable was a key factor in the examination of all data. Restricted cube plots (RCS) were utilized to analyze the strength of the relationship between LAP, CMI, and MetS. The study delved into potential connections between LAP and CMI with MetS, all-cause and cardiovascular mortality using various statistical models such as multivariate logistic regression and Cox regression.

Results

The findings revealed a significant nonlinear association between CMI, LAP, and MetS (P-non-linear < 0.001), irrespective of gender, with all models exhibiting a J-shaped trend. The multivariable logistic regression analysis considered both LAP and CMI as continuous variables or tertiles, revealing significant associations with MetS in male, female, and general populations (All the P < 0.001). Although males displayed a higher risk of MetS, no gender differences were observed in the area under the curve (AUC) values of LAP and CMI for distinguishing (P > 0.005) MetS. Impressively, LAP and CMI were identified as the primary predictors of MetS in both genders from AUC (P < 0.005). More specifically, the cutoff points for distinguishing MetS in females were LAP = 49.87 or CMI = 0.56, while for males, they were LAP = 52.76 or CMI = 0.70. Additionally, the Cox regression analysis revealed that LAP and CMI were correlated with all-cause mortality in both general population and females (P < 0.005), but not in males.

Conclusion

In comparison to other measures of obesity, LAP and CMI demonstrated superior diagnostic accuracy for MetS in both males and females. Additionally, LAP and CMI were found to be predictive of all-cause mortality in both general population and females. These markers are cost-effective, easily accessible, and widely applicable for the early identification and screening of MetS in clinical settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12986-024-00864-2.

Introduction

Cardiovascular disease (CVD) is a significant cause of mortality worldwide, responsible for about 17.9 million deaths in 2019, making up 32% of total global mortality [1]. Metabolic syndrome (MetS) poses a substantial global health challenge, characterized by metabolic issues such as hypertension, obesity, dyslipidemia, and impaired glucose tolerance, all of which are well-known as significant risk factors for CVD [2–6]. Studies have indicated that timely intervention in MetS can effectively reduce the risk of CVD [7–9]. Therefore, timely detection and diagnosis of MetS are essential in the prevention and delay of CVD development, ultimately decreasing CVD-related deaths.

MetS was first introduced by the World Health Organization in 1998 [10] and was frequently associated with obesity and insulin resistance (IR) [11–14]. Recent research suggests that body fat distribution can serve as a substantial predictor of MetS and its related risk factors [11, 15, 16]. Currently, methods used to assess fat distribution in clinical settings, such as computed tomography (CT) scans and magnetic resonance imaging (MRI), tend to be expensive and time-consuming. Thus, it is crucial to identify a straightforward and dependable obesity indicator that can effectively predict MetS. Several epidemiological studies have demonstrated that obesity indicators like body mass index (BMI), waist circumference (WC), and waist-to-height ratio (WHtR) are useful in distinguishing MetS [16–21]. Nevertheless, these indicators only reflect total body fat or subcutaneous adipose tissue (SAT), not visceral adipose tissue (VAT) and there remains ongoing discussion regarding the most suitable indicator for forecasting MetS, along with limited evidence concerning the long-term prognosis of MetS.

The obesity indicators known as lipid accumulation products (LAP) and cardiometabolic index (CMI) have recently attracted considerable interest due to their correlations with obesity and IR [22, 23]. LAP, which integrates triglycerides (TG) and WC, provides a gender-specific assessment of VAT, demonstrating a stronger association with metabolic status. CMI, as a novel index of visceral adiposity, defines obesity through the ratio of TG to high-density lipoprotein cholesterol (HDL-C) in conjunction with WHtR [22, 23]. This study combines body and blood lipid indicators to describe the anatomical and physiological changes associated with VAT. These two measurements serve as indicators of the body’s capacity to store fat reserves. Consequently, elevated levels of LAP and CMI may signify an excess of lipids in ectopic tissues, including the liver, skeletal muscle, heart, blood vessels, kidneys, and pancreas, which are collectively referred to as visceral fat deposition [22, 23]. Numerous studies have shown LAP to be an effective predictor of cardiovascular metabolic conditions, such as hypertension, stroke, diabetes, and psoriasis [24–28]. LAP was also a better predictor of MetS than the triglyceride-glucose (Tyg) index, BMI, WC, WHtR, waist-to-hip ratio (WHR), body roundness index (BRI), a body shape index (ABSI), body adiposity index (BAI), conicity index (CI), abdominal volume index (AVI) visceral obesity index (VAI), and waist triglyceride index (WTI) [29, 30]. Initially, CMI was employed to predict diabetes, but recent studies have linked it to risks of CVD like hypertension and coronary heart disease [28, 31]. Nonetheless, there is limited data on CMI distinguishing MetS, particularly in relation to gender differences. Given the marked gender disparities in the distinguishing power of LAP and CMI for diabetes mellitus and non-alcoholic fatty liver disease (NAFLD), and the varying prevalence of MetS between genders [15, 29], we propose that the predictive capacity of LAP and CMI for MetS may differ by gender. Limited research has delved into how gender differences influence the relationship between obesity-related indicators and MetS. To fill this knowledge gap, this study analyzed data from participants in the National Health and Nutrition Examination Survey (NHANES), aiming to discern the differential roles of LAP and CMI in distinguishing MetS across the general population and specifically among males and females. Additionally, this study evaluated the predictive performance of LAP, CMI, and other obesity-related measures in forecasting MetS.

Method

Study population

NHANES is a comprehensive nationwide study conducted by the National Center for Health Statistics (NCHS), a division of the Centers for Disease Control and Prevention (CDC), with the purpose of evaluating the health and dietary status of individuals in the United States(https://wwwn.cdc.gov/nchs/nhanes/). The NHANES protocol has been formally approved by the CDC’s Institutional Review Board, and all participants give voluntary informed consent. Data was collected through interviews and examinations carried out by well-trained technicians. To ensure a nationally representative sample is obtained every two years, the NHANES survey utilizes a complex multistage, stratified, cluster probability sampling method. For this particular study, this study analyzed a total of 60,936 individuals from the NHANES database spanning the years 2005 to 2018. Subsequently, participants were excluded based on specific criteria: (1) being under the age of 20; (2) having incomplete data required for calculating the LAP and CMI; (3) lacking necessary data for evaluating MetS; (4) missing mortality data, resulting in a final cohort of 11,687 eligible individuals (Fig. 1). MetS was determined in accordance with the NCEP-ATP III guidelines [33], which require at least three of the following five criteria for diagnosis: (1) abdominal obesity (WC ≥ 102 cm in males and WC ≥ 88 cm in females); (2) elevated TG levels ≥ 150 mg/dL; (3) low HDL-C levels < 1.03 mmol/L in males and < 1.29 mmol/L in females; (4) systolic blood pressure (SBP) ≥ 130 mmHg, diastolic blood pressure (DBP) ≥ 85 mmHg or history of hypertension treatment; and (5) elevated fasting blood glucose (FBG) levels ≥ 5.6 mmol/L or previous diagnosis of diabetes mellitus type 2 (T2DM).

Fig. 1.

Flow chart for the enrollment of study population. CMI, cardio-metabolic index; LAP, lipid accumulation product; MetS, metabolic syndrome

Exposure variables

Basic demographic information of participants, inclusive of gender, age, educational level, ethnicity, household income to poverty ratio (PIR), alcohol and tobacco use, as well as medical history encompassing hypertension and diabetes, was obtained through interviews conducted at their homes. Physical measurements, like height, weight, WC, and blood pressure, were taken at a mobile examination facility. Furthermore, laboratory assessments were conducted to analyze indicators such as FBG, total cholesterol (TC), TG, HDL-C, low-density lipoprotein cholesterol (LDL-C), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and glycated hemoglobin A1C (HbA1C) levels, following established procedures.

According to Wakabayashi et al. [32], CMI was calculated based on the physical examination data and blood biochemistry data, the computational formula for CMI was presented as follows:

WHtR = WC (cm)/height (cm) [29].

CMI = [TG (mmol/L)/HDL-c (mmol/L)] × WHtR.

The LAP calculation formula is adjusted according to gender:

LAP = TG (mmol/L) × [WC (cm)-58] for females and LAP = TG (mmol/L) × [WC (cm)-65] for males. TC to HDL ratio (TC/HDL) = TC (mg/dl) /HDL-C(mg/dl), and LDL to HDL ratio (LDL/HDL) = LDL-C(mg/dl) /HDL-C(mg/dl) [22]; Hypertension was described as having a SBP of 140 mmHg or higher, aDBP of 90 mmHg or higher, currently taking medication for hypertension, or a previous diagnosis of hypertension [33]. Diabetes criteria included a FBG level of 7.0 mmol/L or higher, a two-hour plasma glucose level of 11.1 mmol/L or higher after a 75-g oral glucose tolerance test (OGTT), an HbA1c of over 6.5%, or a previous diagnosis of T2DM [33]. Alcohol consumption was determined by consuming at least one drink per month in the past year.

Outcomes

The primary outcome of this study is MetS, while the secondary outcomes include all-cause mortality and cardiovascular mortality. The NCHS maintains the National Death Statistics Database, a centralized repository that combines data from population surveys conducted by the National Population Statistics Center with death certificate information from the National Death Database. While prioritizing participant confidentiality, this database releases data on adult participant mortality. The publicly accessible Linked Mortality Files (LMF) provide mortality follow-up information from the time of survey participation until December 31, 2019, representing the most recent data available (https://www.cdc.gov/nchs/data-linkage). Based on the ICD-10 coding system, causes of death for participants include heart diseases, cancer, respiratory diseases, accidents, strokes, Alzheimer’s, diabetes, influenza, kidney diseases, and other miscellaneous causes. This study defines all-cause mortality as deaths resulting from various factors (including the mentioned 10 causes), and cardiovascular mortality as fatalities linked to cardiovascular-related diseases.

Statistical analysis

Using R4.3.0, this study conducted the statistical analysis and imputed missing data through the random forest method. To address the complex multi-stage survey design with unequal individual sampling, sample weights were incorporated during data analysis to account for differential selection probabilities, non-coverage, and non-response. Analysis was stratified by sex, non-normally distributed continuous variables underwent transformations like logarithmic conversion to achieve normality. Categorical variables were represented as percentages, while continuous variables were stated as mean ± standard deviation (SD). Weighted two-sample t-tests and chi-square tests were used to evaluate differences between groups. Restricted cubic spline (RCS) regression was used to explore the nonlinear relationship between LAP, CMI and MetS. Likelihood ratio test was used to test nonlinearity. Multicollinearity among independent variables was assessed using the variance inflation factor (VIF) method. Covariates with a VIF > 5 were iteratively removed until all remaining variables had a VIF < 5. The Schoenfeld residuals test was applied to assess the proportional hazards (PH) assumption. LAP and CMI were categorized into tertiles (T1, T2, T3), and weighted logistic and Cox regression models were used to examine their associations with MetS, all-cause mortality, and cardiovascular mortality. Model 1 contained LAP or CMI. In the male (female) population, model 2 was adjusted age, race, education, and PIR. In the overall population, gender was added as an additional covariate to Model 2. Model 3 was adjusted for FBG, HbA1c, TC, AST, ALT, HDL-C, smoke, alcohol, SBP and DBP based on Model2. Results were reported as odds ratios (ORs) or hazard ratios (HRs) with corresponding 95% confidence intervals (CIs). Kaplan-Meier curves depicted survival probabilities, and receiver operating characteristic (ROC) curves with the area under the curve (AUC) values compared the performance of CMI, LAP, and other obesity indices (WC, BMI, WHtR, TC/HDL, LDL/HDL) in identifying MetS. Additionally, to validate the stability of the results, we conducted a sensitivity analysis. First, we re-evaluated the relationships between LAP, CMI, MetS, all-cause mortality, and cardiovascular mortality after excluding data with missing values. Second, we performed multiple imputations, conducting 10 iterations to create 10 complete datasets, and repeatedly verified the associations between LAP, CMI, and MetS within these datasets. A statistically significant result was defined as P < 0.05.

Results

Basic characteristics of participants

Table 1 displayed the basic characteristics of all individuals in the study, while Table 2 showcased similar data sorted by sex. A total of 11,687 suitable individuals took part in this research (5,964 females and 5,723 males), with a median age of 49.74 years. As per diagnostic standards, 2,542 (42.62%) females and 2,350(41.06%)males were diagnosed with MetS. In general, in comparison to individuals without MetS, those with MetS were generally older, had lower income, less education, and more pre-existing conditions (P < 0.05). Additionally, irrespective of sex, individuals with MetS showed notably raised levels of WC, BMI, DBP, SBP, FBG, HbA1c, TG, TC, ALT, AST, LAP, CMI, LDL-C/HDL-C, TC/HDL-C, and WHtR while having decreased HDL-C levels in contrast to those without MetS (P < 0.05). Among females, LDL-C levels were notably higher in those with MetS compared to those without MetS (P < 0.05). Among males, the percentage of current alcohol consumption was significantly greater in those with MetS compared to those without MetS (P < 0.05).

Table 1.

Baseline characteristics of subjects stratified by MetS in general population

Variable	Overall	Non–MetS	MetS	P value
	(n=11687)	(n=6795)	(n=4892)
Male (%)	5723(49.0)	3373 (49.6)	2350 (48.0)	<0.001a
Race (%)				<0.001a
Mexican American	1819 (15.6)	1002 (14.7)	817 (16.7)
Non-Hispanic White	5019 (42.9)	2804 (41.3)	2215 (45.3)
Non-Hispanic Black	2264 (19.4)	1355 (19.9)	909 (18.6)
Other/multiracial	2585 (22.1)	1634(24.0)	951 (19.4)
Age (years)				<0.001 a
20-40(years)	3810 (32.6)	2931 (43.1)	879 (18.0)
40-60(years)	3981 (34.1)	2251 (33.1)	1730 (35.4)
>60(years)	3896 (33.3)	1613 (23.7)	2283 (46.7)
Education (%)				<0.001 a
Less than high school	3022 (25.9)	1521 (22.4)	1501 (30.7)
High school graduate	2620 (22.4)	1458 (21.5)	1162 (23.8)
Some college or above	6045 (51.7)	3816 (56.2)	2229 (45.6)
PIR	2.46 (1.62)	2.55 (1.65)	2.33 (1.58)	<0.001 b
WC (cm)	99.17 (16.09)	92.01 (13.25)	109.11 (14.31)	<0.001 b
Hight(cm)	167.26 (10.09)	167.57 (9.80)	166.83 (10.47)	<0.001 b
BMI (kg/m2)	28.96 (6.69)	26.38 (5.42)	32.53 (6.64)	<0.001 b
DBP (mmHg)	69.51 (12.82)	68.20 (11.59)	71.32 (14.16)	<0.001 b
SBP (mmHg)	124.04 (18.61)	119.19 (16.91)	130.77 (18.77)	<0.001 b
FBG (mg/dl)	109.49 (36.49)	98.91 (21.37)	124.19 (46.63)	<0.001 b
HbA1c	5.78 (1.11)	5.47 (0.69)	6.21 (1.40)	<0.001 b
Smoke (%)	5277 (45.2)	2859 (42.1)	2418 (49.4)	<0.001 a
Drinking (%)	8408 (71.9)	5053 (74.4)	3355 (68.6)	<0.001 a
HDL-C (mg/dl)	53.92 (16.19)	58.92 (15.75)	46.99 (14.10)	<0.001 b
LDL-C (mg/dl)	114.17 (35.52)	112.69 (34.12)	116.21 (37.29)	<0.001 b
TG (mg/dl)	130.86 (126.57)	94.55 (53.04)	181.29 (173.18)	<0.001 b
TC (mg/dl)	193.48 (42.19)	191.26 (39.24)	196.57 (45.80)	<0.001 b
AST (u/l)	25.55 (20.14)	23.78 (19.76)	28.01 (20.42)	<0.001 b
ALT (u/l)	26.05 (21.77)	25.32 (20.79)	27.07 (23.04)	0.001 b
CMI	0.80 (1.26)	0.43 (0.35)	1.30 (1.79)	<0.001 b
LAP	59.42 (71.60)	33.31 (24.01)	95.67 (95.85)	<0.001 b
WHtR	0.59 (0.10)	0.55 (0.08)	0.66 (0.08)	<0.001 b
LDL-C/HDL-C	2.31 (1.03)	2.05 (0.83)	2.66 (1.16)	<0.001 b
TC/HDL-C	3.87 (1.42)	3.42 (0.99)	4.49 (1.67)	<0.001 b
T2DM (%)	6449 (55.2)	2195 (32.3)	4254 (87.0)	<0.001 a
Hypertension (%)	6021 (51.5)	2096 (30.8)	3925 (80.2)	<0.001 a

PIR, poverty index ratio; WC, waist circumference; BMI, body mass index; DBP, Diastolic blood pressure; SBP, Systolic blood pressure; FBG, fasting blood glucose; HbA1c, Glycosylated Hemoglobin type A1C; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, triglycerides; TC, total cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CMI, cardio-metabolic index; LAP, lipid accumulation product; WHtR, Waist-to-Height Ratio; T2DM, diabetes mellitus type 2; MetS, Metabolic syndrome. ^a, categorical variables were compared using the chi-square test;^b, continuous variables were compared using the two-sample t-test

Table 2.

Baseline characteristics of subjects stratified by gender

Variables	Female (n = 5964)		P value	Male (n = 5723)		P value
	Non–MetS (n=3422)	MetS (n=2542)		Non–MetS (n= 3373)	MetS (n=2350)
Race (%)			<0.001a			<0.001a
Mexican American	502 (14.7)	431 (17.0)		500 (14.8)	386 (16.4)
Non-Hispanic White	1440 (42.1)	1068 (42.0)		1364 (40.4)	1147 (48.8)
Non-Hispanic Black	652 (19.1)	534 (21.0)		703 (20.8)	375 (16.0)
Other/multiracial	828 (24.2)	509 (20.0)		806 (23.9)	442 (18.8)
Age (%)			<0.001 a			<0.001 a
20-40(years)	1530 (44.7)	416 (16.4)		1401 (41.5)	463 (19.7)
40-60(years)	1148 (33.5)	888 (34.9)		1103 (32.7)	842 (35.8)
>60(years)	744 (21.7)	1238 (48.7)		869 (25.8)	1045 (44.5)
Education (%)			<0.001 a			0.007 a
Less than high school	663 (19.4)	828 (32.6)		858 (25.4)	673 (28.6)
High school graduate	661 (19.3)	590 (23.2)		797 (23.6)	572 (24.3)
Some college or above	2098 (61.3)	1124 (44.2)		1718 (50.9)	1105 (47.0)
PIR	2.53 (1.66)	2.14 (1.51)	<0.001b	2.57 (1.63)	2.54 (1.62)	0.552 b
WC (cm)	90.14 (13.86)	107.07 (14.46)	<0.001b	93.91 (12.31)	111.33 (13.81)	<0.001 b
Hight(cm)	161.14 (7.04)	159.77 (7.24)	<0.001b	174.09 (7.66)	174.46 (7.70)	0.074 b
BMI (kg/m2)	26.58 (6.09)	33.09 (7.25)	<0.001b	26.18 (4.64)	31.93 (5.84)	<0.001 b
DBP (mmHg)	66.68 (10.91)	69.59 (13.83)	<0.001 b	69.75 (12.05)	73.20 (14.28)	<0.001 b
SBP (mmHg)	116.16 (16.90)	130.23 (20.01)	<0.001 b	122.27 (16.35)	131.36 (17.32)	<0.001 b
FBG (mg/dl)	94.97 (16.72)	121.93 (45.57)	<0.001 b	102.91 (24.60)	126.64 (47.64)	<0.001 b
HbA1c	5.27 (0.93)	6.77 (2.53)	<0.001 b	5.55 (0.80)	6.21 (1.45)	<0.001 b
Smoke (%)	1116 (32.6)	1031 (40.6)	<0.001a	1743 (51.7)	1387 (59.0)	<0.001 a
Drinking (%)	2238 (65.4)	1392 (54.8)	<0.001a	2815 (83.5)	1963 (83.5)	0.969 a
HDL-C (mg/dl)	64.15 (15.18)	51.43 (14.65)	<0.001 b	53.60 (14.49)	42.19 (11.74)	<0.001 b
LDL-C (mg/dl)	111.12 (33.64)	118.00 (37.03)	<0.001 b	114.29 (34.53)	114.28 (37.49)	0.995 b
TG (mg/dl)	87.32 (41.87)	165.14(167.44)	<0.001 b	101.88 (61.51)	198.76 (177.57)	0.001 b
TC (mg/dl)	193.77(39.41)	200.49 (44.85)	<0.001 b	188.72 (38.92)	192.33 (46.45)	0.001 b
AST (u/l)	20.08 (19.18)	24.06 (17.74)	<0.001 b	27.53 (19.63)	32.28 (22.19)	<0.001 b
ALT (u/l)	23.25 (20.79)	25.48 (22.02)	<0.001 b	27.42 (20.58)	28.79 (23.98)	0.022 b
CMI	0.37 (0.23)	1.09 (1.65)	<0.001 b	0.51 (0.42)	1.53 (1.90)	<0.001 b
LAP	32.69 (23.09)	91.04 (102.27)	<0.001 b	33.94 (24.90)	100.68 (88.13)	<0.001 b
WHtR	0.54 (0.07)	0.64 (0.08)	<0.001 b	0.56 (0.09)	0.67 (0.09)	<0.001 b
LDL-C/HDL-C	1.83 (0.70)	2.45 (1.01)	<0.001 b	2.28 (0.89)	2.89 (1.28)	<0.001 b
TC/HDL-C	3.14 (0.80)	4.14 (1.39)	<0.001 b	3.71 (1.08)	4.87 (1.85)	<0.001 b
T2DM (%)	693 (20.3)	2148 (84.5)	<0.001a	1502 (44.5)	2106 (89.6)	<0.001 a
Hypertension (%)	892 (26.1)	2006 (78.9)	<0.001a	1204 (35.7)	1919 (81.7)	<0.001 a

PIR, poverty index ratio; WC, waist circumference; BMI, body mass index; DBP, Diastolic blood pressure; SBP, Systolic blood pressure; FBG, fasting blood glucose; HbA1c, Glycosylated Hemoglobin type A1C; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, triglycerides; TC, total cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CMI, cardio-metabolic index; LAP, lipid accumulation product; WHtR, Waist-to-Height Ratio; T2DM, diabetes mellitus type 2; MetS, Metabolic syndrome. ^a, categorical variables were compared using the chi-square test; ^b, continuous variables were compared using the two-sample t-test

Trend relationship of LAP and CMI with the risk of MetS

Utilizing RCS, this study investigated the dose-response correlation between LAP, CMI, and MetS (Fig. 2). These findings revealed a significant nonlinear association between LAP, CMI and MetS(P-non-linear < 0.001), irrespective of gender, with all models exhibiting a J-shaped trend.

Fig. 2.

Association between LAP, CMI and metabolic syndrome (MetS). The figure consists of two parts: a frequency distribution histogram and a line graph. The histogram displays the distribution of individuals across different log(LAP) or log(CMI) levels, while the line graph illustrates the odds ratios (OR) for the risk of Mets at varying log(LAP) or log(CMI) concentrations. The x-axis represents log(LAP) or log(CMI) levels, the left y-axis corresponds to OR values, and the right y-axis indicates the number of individuals. (A) Association between LAP and MetS in general population; (B) Association between LAP and MetS in female; (C) Association between LAP and MetS in male; (D) Association between CMI and MetS in general population; (E) Association between CMI and MetS in female; (F) Association between CMI and MetS in male. CMI, cardio-metabolic index; LAP, lipid accumulation product; MetS, metabolic syndrome. The p-value for nonlinearity was calculated using the Likelihood ratio test.

Influence of LAP and CMI on MetS detection through multivariate logistic regression

The VIF values for WC, height, TG, and LDL were greater than 5, so they were excluded from the subsequent logistic models. The findings from the multivariate logistic regression analysis can be seen in Table 3, where LAP and CMI were included as continuous variables or divided into tertiles. It was observed that both LAP and CMI showed significant associations with MetS, regardless of gender.

Table 3.

Multivariate logistic regression of LAP and CMI for MetS

	Model1		Model2		Model3
	OR (95%CI)	Pa value	OR (95%CI)	Pa value	OR (95%CI)	Pa value
General population
ln-transformed LAP	15.80(14.00, 17.90)	<0.001	12.80(10.7, 15.20)	<0.001	11.30(8.60, 14.70)	<0.001
ln-transformed LAP levels
T1(<3.38)	Ref		Ref		Ref
T2(3.38,4.10)	8.20(6.67, 10.10)	<0.001	3.85(3.06, 4.86)	<0.001	3.92(3.09, 4.98)	<0.001
T3(>4.10)	72.40(56.7, 92.30)	<0.001	19.8(15.00, 26.20)	<0.001	20.30(15.10,27.30)	<0.001
P value for trend		<0.001		<0.001		<0.001
ln-transformed CMI	9.82(8.85, 10.90)	<0.001	8.76(7.78, 9.86)	<0.001	10.30(7.77,13.70)	<0.001
ln-transformed CMI levels
T1(<-1.03)	Ref		Ref		Ref
T2(-1.03, -0.31)	4.04(3.31, 4.92)	<0.001	2.63 (2.19, 3.16)	<0.001	1.87(1.52, 2.28)	<0.001
T3(>-0.31)	34.80(29.10, 41.70)	<0.001	23.80(19.30, 29.20)	<0.001	12.60(8.62, 18.40)	<0.001
P value for trend		<0.001		<0.001		<0.001
Females
ln-transformed LAP	15.30(12.80, 18.30)	<0.001	12.20(9.62, 15.50)	<0.001	9.29(6.90, 12.50)	<0.001
ln-transformed LAP levels
T1(< 3.36)	Ref		Ref		Ref
T2(3.36, 4.09)	7.97(6.16, 10.30)	<0.001	4.69(3.58, 6.14)	<0.001	3.07(2.31, 4.07)	<0.001
T3(>4.09)	69.00(50.40, 94.50)	<0.001	32.3(22.7, 45.90)	<0.001	14.80(10.20,21.40)	<0.001
P value for trend		<0.001		<0.001		<0.001
ln-transformed CMI	12.30(10.30, 14.60)	<0.001	10.20(8.26,12.50)	<0.001	9.50(6.58, 13.70)	<0.001
ln-transformed CMI levels
T1(< -1.16)	Ref		Ref		Ref
T2(-1.16, -0.45)	4.60(3.08, 6.89)	<0.001	2.75(1.93, 3.92)	<0.001	1.44(1.02, 2.02)	0.038
T3(> -0.45)	38.50(27.70, 53.40)	<0.001	21.30(14.70, 30.90)	<0.001	7.30(4.26, 12.50)	<0.001
P value for trend		<0.001		<0.001		<0.001
Males
ln-transformed LAP	16.60(14.20, 19.50)	<0.001	13.20(10.90, 16.10)	<0.001	11.50(8.22, 16.00)	<0.001
ln-transformed LAP levels
T1(< 3.41)	Ref		Ref		Ref
T2(3.41,4.11)	9.03(6.22, 13.10)	<0.001	5.10(3.47, 7.51)	<0.001	4.82(2.84, 8.17)	<0.001
T3(>4.11)	82.80(59.20, 116)	<0.001	39.0(27.80, 54.90)	<0.001	24.1(14.40, 40.30)	<0.001
P value for trend		<0.001		<0.001		<0.001
ln-transformed CMI	9.91(8.56, 11.50)	<0.001	9.26(7.83, 11.00)	<0.001	10.40(7.22, 14.90)	<0.001
ln-transformed CMI levels
T1(< -0.89)	Ref		Ref		Ref
T2(-0.89, -0.17)	3.35(2.72, 4.13)	<0.001	2.33(1.730, 3.13)	<0.001	1.46(1.00, 2.14)	0.052
T3(> -0.17)	31.80(25.60, 39.50)	<0.001	24.60(18.10, 33.50)	<0.001	11.50(6.75, 19.70)	<0.001
P value for trend		<0.001		<0.001		<0.001

Model1 = LAP/CMI

Model2 = Model 1+ age, race, education, PIR (Sex was added as an additional covariate in the general population)

Model3 = Model 2+ FBG, HbA1c, TC, AST, ALT, HDL-C, smoke, alcohol, SBP, DBP

T, Tertile; Ref, reference; OR, odds ratio; CI, Confidence Interval; LAP, lipid accumulation product; CMI, the Cardiometabolic Index; PIR, poverty index ratio; DBP, Diastolic blood pressure; SBP, Systolic blood pressure; FBG, fasting blood glucose; HbA1c, Glycosylated Hemoglobin type A1C; HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; the Cardiometabolic Index (CMI); Metabolic syndrome (MetS)

^a, logistic regression calculates the P-value for each variable using the Wald test

In the general population, these associations were statistically significant after adjusting for potential confounders [ln-transformed LAP: OR = 11.30 (95% CI = 8.60–14.70); ln-transformed CMI: OR = 10.30 (95% CI = 7.77–13.70) per 1 unit increase; all P < 0.001]. Further analysis indicated that the impact of elevated ln-transformed LAP and CMI on the risk of MetS was more prominent in males compared to females. Adjusting for potential confounders (Model 3) revealed in females [ln-transformed LAP: OR = 9.29 (95% CI = 6.90–12.50); ln-transformed CMI: OR = 9.50 (95% CI = 6.58–13.7) per 1 unit increase; all P < 0.001], and in males [ln-transformed LAP: OR = 11.50 (95% CI = 8.22-16.00); ln-transformed CMI: OR = 10.40 (95% CI = 7.22–14.90) per 1 unit increase; all P < 0.001] (Table 3).

Even when LAP and CMI were categorized into tertiles, their association with MetS remained statistically significant. Adjusting for potential confounders (Model 3), individuals in the highest tertile of LAP in the general population had a 20.3 times higher risk of MetS (95% CI = 15.10–27.30), with females at 14.80 times (95% CI = 10.20–21.40) and males at 24.1 times (95% CI = 14.43–40.30). Similarly, participants in the highest CMI tertile had a 12.60 times higher risk of MetS in the general population (95% CI = 15.10–27.30), with females at 7.30 times (95% CI = 4.26–12.50) and males at 11.50 times (95% CI = 6.75–19.70). Notably, irrespective of gender, as the tertiles of LAP and CMI increased, there was a gradual rise in the prevalence of MetS (all trends p < 0.001) (Table 3).

Association of LAP and CMI with all-cause and cardiovascular mortality

The multivariable Cox regression analysis included LAP and CMI as either continuous variables or grouped into tertiles, showing significant gender differences in their connection to all-cause mortality (Table 4). Both LAP and CMI were linked to an increased risk of all-cause mortality in the overall population (P < 0.05). Even after adjusting for potential confounding variables, this relationship remained significant [ln-transformed LAP, HR = 1.18 (95%CI: 1.01–1.37); ln-transformed CMI, HR = 1.28 (95%CI: 1.10–1.48), all P-values < 0.05]. The statistical significance of LAP and CMI in relation to all-cause mortality persisted when examined by tertiles (P < 0.05).

Table 4.

Multiple Cox regression of LAP and CMI for all-cause mortality

	Model1		Model2		Model3
	HR (95%CI)	Pa value	HR (95%CI)	Pa value	HR (95%CI)	Pa value
General population
ln-transformed LAP	1.37(1.29, 1.47)	<0.001	1.14(1.03, 1.26	0.012	1.18(1.01, 1.37)	0.039
LAP levels
T1(<3.38)	Ref		Ref		Ref
T2(3.38,4.10)	1.54(1.27, 1.86)	<0.001	1.08(0.88, 1.31)	0.500	1.09(0.89, 1.33)	0.400
T3(>4.10)	1.96(1.64, 2.33)	<0.001	1.29(1.04, 1.59)	0.019	1.32(1.03, 1.68)	0.025
P value for trend		<0.001		<0.001		<0.001
ln-transformed CMI	1.29(1.21, 1.37)	<0.001	1.19(1.09, 1.29)	<0.001	1.28(1.10, 1.48)	0.001
CMI levels
T1(<-1.03)	Ref		Ref		Ref
T2(-1.03, -0.31)	1.37(1.10, 1.71)	0.004	1.22(0.97, 1.52)	0.093	1.29(1.05, 1.58)	0.016
T3(>-0.31)	1.80(1.54, 2.11)	<0.001	1.48(1.25, 1.75)	<0.001	1.61(1.27, 2.03)	<0.001
P value for trend		<0.001		<0.001		<0.001
Females
ln-transformed LAP	1.54(1.36, 1.73)	<0.001	1.28(1.07, 1.53)	0.006	1.46(1.14, 1.87)	0.003
LAP levels
T1(< 3.36)	Ref		Ref		Ref
T2(3.36, 4.09)	1.58(1.17, 2.13)	0.003	1.04(0.78, 1.39)	0.800	1.06(0.80, 1.41)	0.700
T3(>4.09)	2.27(1.68, 3.08)	<0.001	1.38(1.00, 1.89)	0.050	1.48(0.97, 2.25)	0.069
P value for trend		<0.001		<0.001		0.070
ln-transformed CMI	1.51(1.34, 1.71)	<0.001	1.31(1.15, 1.51)	<0.001	1.81(1.42, 2.31)	<0.001
CMI levels
T1(< -1.16)	Ref		Ref		Ref
T2(-1.16, -0.45)	1.35(0.96, 1.91)	0.085	1.06(0.76, 1.48)	0.700	1.26(0.84, 1.89)	0.300
T3(> -0.45)	2.35(1.74, 3.16)	<0.001	1.64(1.24, 2.16)	<0.001	2.29(1.48, 3.53)	<0.001
P value for trend		<0.001		<0.001		<0.001
Males
ln-transformed LAP	1.24(1.11, 1.40)	<0.001	1.01(0.85, 1.20)	0.900	1.10(0.90, 1.35)	0.300
LAP levels
T1(< 3.41)	Ref		Ref		Ref
T2(3.41,4.11)	1.33(1.01, 1.75)	0.042	0.92(0.71, 1.19)	0.500	0.97(0.72, 1.29)	0.800
T3(>4.11)	1.58(1.21, 2.04)	<0.001	1.03(0.75, 1.42)	0.800	1.14(0.76, 1.70)	0.500
P value for trend		0.001		<0.001		0.483
ln-transformed CMI	1.10(0.97, 1.24)	0.14	1.01(0.87, 1.17)	0.900	1.09(0.88, 1.35)	0.400
CMI levels
T1(< -0.89)	Ref		Ref		Ref
T2(-0.89, -0.17)	1.10(0.80, 1.52)	0.500	0.93(0.68, 1.28)	0.700	0.98(0.69, 1.40)	>0.900
T3(> -0.17)	1.25(0.97, 1.62)	0.079	1.06(0.79, 1.43)	0.700	1.15(0.78, 1.69)	0.500
P value for trend		0.078		<0.001		0.297

Model1 = LAP/CMI

Model2 = Model 1+ age, race, education, PIR (Sex was added as an additional covariate in the general population)

Model3 = Model 2+ FBG, HbA1c, TC, AST, ALT, HDL-C, smoke, alcohol, SBP, DBP

T, Tertile; Ref, reference; HRs, hazard ratios; CI, Confidence Interval; LAP, lipid accumulation product; CMI, the Cardiometabolic Index; PIR, poverty index ratio; DBP, Diastolic blood pressure; SBP, Systolic blood pressure; FBG, fasting blood glucose; HbA1c, Glycosylated Hemoglobin type A1C; HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CMI, the Cardiometabolic Index; MetS, Metabolic syndrome

^a, Cox regression calculates the P-value for each variable using the Wald test

Among females, the association between LAP, CMI, and all-cause mortality was notably stronger [ln-transformed LAP, HR = 1.54 (95% CI: 1.36–1.73); ln-transformed CMI, HR = 1.51 (95% CI: 1.34–1.71); all P-values < 0.001]. Upon adjusting for potential confounders, these relationships remained significant, with respective HRs for ln-transformed LAP and CMI at 1.46 (95% CI: 1.14–1.87) and 1.81 (95% CI: 1.42–2.31), all P-values < 0.001. Conversely, no significant increase in the risk of all-cause mortality was observed with LAP and CMI in the male population (P > 0.05). The comparison of all-cause mortality rates across varying levels of LAP and CMI in the general, female and male population is depicted in Fig. 3, showing a noticeable increase as the tertile levels of LAP and CMI increased (log-rank test: P < 0.001). The analysis of LAP, CMI and cardiovascular mortality is summarized in Table 5. Upon adjusting for confounding factors, no significant association between LAP, CMI, and cardiovascular mortality was evident in the general, female, and male population (P > 0.05).

Fig. 3.

The Kaplan-Meier analysis of the prognostic effect of different levels of LAP, CMI on all-cause mortality and cardiovascular (CV) mortality (A) The Kaplan-Meier analysis of the prognostic effect of different levels of LAP on all-cause mortality in general population. (B) The Kaplan-Meier analysis of the prognostic effect of different levels of LAP on all-cause mortality in female. (C) The Kaplan-Meier analysis of the prognostic effect of different levels of LAP on all-cause mortality in male. (D) The Kaplan-Meier analysis of the prognostic effect of different levels of CMI on all-cause mortality in general population. (E) The Kaplan-Meier analysis of the prognostic effect of different levels of CMI on all-cause mortality in female. (F) The Kaplan-Meier analysis of the prognostic effect of different levels of CMI on all-cause mortality in male. CMI, cardio-metabolic index; LAP, lipid accumulation product; MetS, metabolic syndrome. The p-value for Kaplan-Meier (KM) survival curves was calculated using the log-rank test

Table 5.

Multiple Cox regression of LAP and CMI for cardiovascular mortality

	Model1		Model2		Model3
	HR (95%CI)	P a value	HR (95%CI)	P a value	HR (95%CI)	Pa value
General population
ln-transformed LAP	1.50(1.26, 1.77)	<0.001	1.17(0.89, 1.55)	0.300	1.25(0.82, 1.91)	0.300
LAP levels
T1(<3.38)	Ref		Ref		Ref
T2(3.38,4.10)	1.96(1.24, 3.08)	0.004	1.24(0.76, 2.03)	0.400	1.19(0.69, 2.06)	0.500
T3(>4.10)	2.36(1.59, 3.50)	<0.001	1.30(0.87, 1.95)	0.200	1.29(0.80, 2.09)	0.300
ln-transformed CMI	1.35(1.16, 1.56)	<0.001	1.20(0.99, 1.46)	0.060	1.23(0.84, 1.81)	0.300
CMI levels
T1(<-1.03)	Ref		Ref		Ref
T2(-1.03, -0.31)	1.83(1.19, 2.81)	0.006	1.49(0.95, 2.33)	0.085	1.29(0.77, 2.15)	0.300
T3(>-0.31)	1.83(1.26, 2.67)	0.002	1.30(0.87, 1.94)	0.200	1.07(0.62, 1.84)	0.800
Females
ln-transformed LAP	1.59(1.27, 1.99)	<0.001	1.17(0.83, 1.66)	0.400	1.35(0.76, 2.40)	0.300
LAP levels
T1(< 3.36)	Ref		Ref		Ref
T2(3.36, 4.09)	1.98(1.07, 3.67)	0.030	1.09(0.61, 1.93)	0.800	1.04(0.51, 2.13)	>0.900
T3(>4.09)	2.40(1.22, 4.71)	0.011	1.08(0.54, 2.18)	0.800	1.09(0.40, 2.99)	0.900
ln-transformed CMI	1.53(1.26, 1.86)	<0.001	1.25(0.96, 1.64)	0.100	1.61(0.94, 2.75)	0.085
CMI levels
T1(< -1.16)	Ref		Ref		Ref
T2(-1.16, -0.45)	1.52(0.83, 2.78)	0.200	1.08(0.63, 1.85)	0.800	1.08(0.54, 2.17)	0.800
T3(> -0.45)	2.52(1.41, 4.50)	0.002	1.54(0.83, 2.84)	0.200	1.71(0.66, 4.41)	0.300
Males
ln-transformed LAP	1.41(1.13, 1.76)	0.003	1.12(0.81, 1.54)	0.500	1.37(0.88, 2.13)	0.200
LAP levels
T1(< 3.41)	Ref		Ref		Ref
T2(3.41,4.11)	1.85(0.74, 4.63)	0.200	1.22(0.48, 3.11)	0.700	1.31(0.48, 3.58)	0.600
T3(>4.11)	2.25(1.09, 4.63)	0.028	1.36(0.60, 3.08)	0.500	1.63(0.60, 4.41)	0.300
ln-transformed CMI	1.17(0.93, 1.47)	0.200	1.04(0.79, 1.36)	0.8000	1.22(0.78, 1.93)	0.400
CMI levels
T1(< -0.89)	Ref		Ref		Ref
T2(-0.89, -0.17)	1.73(0.83, 3.60)	0.140	1.48(0.70, 3.16)	0.300	1.58(0.60, 4.15)	0.400
T3(> -0.17)	1.49(0.77, 2.86)	0.200	1.19(0.59, 2.39)	0.600	1.43(0.50, 4.05)	0.500

Model1 = LAP/CMI. Model2 = Model 1+ age, race, education, PIR (Sex was added as an additional covariate in the general population). Model3 = Model 2+ FBG, HbA1c, TC, AST, ALT, HDL-C, smoke, alcohol, SBP, DBP

T, Tertile; Ref, reference; HRs, hazard ratios; CI, Confidence Interval; LAP, lipid accumulation product; CMI, the Cardiometabolic Index; PIR, poverty index ratio; DBP, Diastolic blood pressure; SBP, Systolic blood pressure; FBG, fasting blood glucose; HbA1c, Glycosylated Hemoglobin type A1C; HDL-C, high-density lipoprotein cholesterol; TG, triglycerides; TC, total cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CMI, the Cardiometabolic Index; MetS, Metabolic syndrome.^a, Cox regression calculates the P-value for each variable using the Wald test

Assessment of obesity-related indicators for distinguishing MetS

We compared the discriminative ability of LAP and CMI for MetS from both cross-sectional and longitudinal perspectives, with a focus on gender differences. In the cross-sectional analysis, we summarized the AUC values of various obesity markers for distinguishing MetS (Table 6; Fig. 4). The longitudinal comparison highlights the performance of different obesity markers in discriminating MetS among males and females (Fig. 5). Notably, LAP demonstrated the strongest discriminative power for MetS across the general population, as well as within female and male groups, followed by CMI. Other obesity markers exhibited significantly weaker discrimination compared to LAP and CMI. Additionally, TC/HDL and LDL/HDL had better discrimination for MetS in females than in males, while WC, WHtR, and BMI performed better in males. Although LAP and CMI showed higher AUC values for discriminating MetS in males compared to females, these differences were not statistically significant (P > 0.05).

Table 6.

AUCs of various indexes for discriminating MetS by sex

Variables	AUC (95%CI)	Cutoff points	Youden’s index	Sensitivity (%)	Specificity (%)
General
LAP	0.881(0.875,0.887)	49.25	0.585	77.37	81.12
CMI	0.849(0.842,0.856)	0.62	0.543	72.67	81.59
WC	0.822(0.815,0.830)	101.75	0.504	69.26	81.16
BMI	0.789(0.780,0.797)	27.26	0.446	80.80	63.80
WHtR	0.833(0.825,0.840)	0.58	0.532	81.58	71.66
TC/HDL	0.722(0.713,0.732)	3.82	0.335	62.90	70.55
LDL/HDL	0.671(0.661,0.681)	2.15	0.254	64.23	61.15
Female
LAP	0.878(0.870,0.887)	49.87	0.582	75.77	82.38
CMI	0.862(0.852,0.871)	0.56	0.559	71.83	84.08
WC	0.814(0.803,0.824)	93.05	0.488	84.66	64.17
BMI	0.776(0.764,0.787)	27.22	0.425	80.25	62.27
WHtR	0.826(0.815,0.853)	0.58	0.517	85.44	66.25
TC/HDL	0.745(0.732,0.758)	3.62	0.388	61.72	77.06
LDL/HDL	0.699(0.685,0.712)	2.07	0.311	61.01	70.05
Male
LAP	0.884(0.876,0.893)	52.76	0.591	75.28	83.84
CMI	0.848(0.838,0.858)	0.70	0.549	74.64	80.25
WC	0.847(0.837,0.857)	101.75	0.604	79.57	80.79
BMI	0.805(0.794,0.817)	27.45	0.471	80.04	67.06
WHtR	0.844(0.834,0.854)	0.58	0.547	78.51	76.22
TC/HDL	0.718(0.704,0.731)	4.38	0.329	56.81	76.05
LDL/HDL	0.656(0.641,0.670)	2.73	0.230	50.85	72.19

CMI, cardio-metabolic index; LAP, lipid accumulation product; WC, waist circumference; BMI, body mass index; WHtR, Waist-to-Height Ratio; TC/HDL, Total cholesterol to HDL ratio; LDL/HDL, LDL to HDL ratio; AUC, area under the curve

Youden’s index= (Sensitivity +Specificity)-1

Fig. 4.

ROC curve comparison of different obesity indicators in prediction of metabolic syndrome (MetS) among (A) General population; (B) Females; (C) Males. CMI, cardio-metabolic index; LAP, lipid accumulation product; WC, waist circumference; BMI, body mass index; WHtR, Waist-to-Height Ratio; TC/HDL, Total cholesterol to HDL ratio; LDL/HDL, LDL to HDL ratio; ROC, receiver operating characteristic; AUC, area under the curve

Fig. 5.

ROC curve comparison of different obesity indicators prediction of metabolic syndrome (MetS) among females and males. (A) LAP; (B) CMI; (C) LDL/HDL; (D) TC/HDL; (E) BMI; (F) WHtR; (G) WC; CMI, cardio-metabolic index; LAP, lipid accumulation product; WC, waist circumference; BMI, body mass index; WHtR, Waist-to-Height Ratio; TC/HDL, Total cholesterol to HDL ratio; LDL/HDL, LDL to HDL ratio; ROC, receiver operating characteristic; AUC, area under the curve. The P-values were calculated using the DeLong test

Sensitivity analyses

In the sensitivity analysis, after excluding participants with missing data, the results remained consistent (Supplementary Tables 2–4). The proportion of missing data was provided in Supplementary Table 1. The findings indicate that both LAP and CMI were significantly associated with MetS in both males and females. Furthermore, LAP and CMI were associated with all-cause mortality in females but not in males. As shown in Supplementary Table 5, the association between LAP, CMI, and MetS was consistent across the 10 imputed datasets, demonstrating the robustness of our results.

Discussion

This research conducted a thorough analysis of the correlation between LAP, CMI, and MetS, as well as the risk of all-cause and cardiovascular mortality. It also examined the effectiveness of seven obesity markers in distinguishing MetS, with the goal of finding a simpler, cheaper, and less invasive method for distinguishing MetS. By controlling for key variables that could influence the results, this study confirmed that LAP and CMI play a crucial role in distinguishing MetS. Regardless of gender, LAP and CMI demonstrated superior distinguishing abilities for MetS in comparison to other obesity indicators like WC, BMI, WHtR, TC/HDL and LDL/HDL. Both LAP and CMI could serve as useful tools for early detection and targeted interventions for MetS. Furthermore, LAP and CMI were linked to higher all-cause mortality rates in the overall population and among females, but did not show any association with cardiovascular mortality.

Recent studies have emphasized LAP and CMI’s improved ability to predict chronic conditions like hypertension, diabetes, coronary heart disease, NAFLD, hyperuricemia, and chronic kidney disease [34–38]. In a comprehensive cross-sectional study in China with 35,446 participants, it was shown that LAP was a more reliable MetS predictor compared to BMI, WC, WHR, and WHtR [39], supporting these findings in this study. Another recent study found that CMI was more effective at distinguishing MetS in women with obesity than WHR, WtHR, and BMI [40]. However, there is a lack of research simultaneously comparing the discriminating accuracy of LAP and CMI for MetS. While Datta et al. demonstrated that LAP and CMI outperformed BMI in distinguishing MetS in a study of 250 Mexican subjects [41], their research was limited by a smaller sample size and only included two obesity markers. In contrast, this study included a wider range of obesity markers like WC, BMI, WHtR, TC/HDL and LDL/HDL and a larger sample size to discern the differential roles of LAP and CMI in distinguishing MetS across the general population and specifically among males and females. Additionally, this study validated LAP and CMI’s distinguishing significance for all-cause mortality, highlighting their enhanced innovation and clinical relevance.

Obesity represents a clinical condition attributed to an imbalance in lipid and glucose metabolism, marked by fat accumulation in both SAT and VAT across various regions of the body [16]. Numerous studies have highlighted the significant role of VAT in metabolic disorders and IR [11]. While BMI is a standard tool for evaluating obesity, its accuracy is restricted as it only delivers a broad approximation of body fat [42]. The specific localization of fat is deemed vital due to its distinct metabolic consequences [16]. Other measurements such as WC and WHtR have been recommended for their superior predictive capability for metabolic diseases over BMI [43]. However, WC and WHtR lack efficacy in differentiating between SAT and VAT. The TC/HDL and LDL/HDL ratios are linked to the risk of atherosclerosis, with smaller, denser LDL particles presenting a greater risk compared to larger particles [44]. BMI, WC, and WHtR merely indicate fat distribution, while TC/HDL and LDL/HDL ratios exclusively represent blood lipid profiles. Recently developed metrics such as LAP and CMI have exhibited superior capability in distinguishing metabolic disorders [36]. LAP combines WC and TG levels, both of which are strongly associated with visceral fat accumulation. CMI includes TG and HDL, capturing the dyslipidemia often linked with excess visceral fat. These markers are particularly valuable as they provide a non-invasive and practical method to assess VAT, which is a significant risk factor for MetS and related diseases [22, 23]. This research indicated that LAP and CMI were more effective in distinguishing MetS compared to other obesity measures, ranking like WC, WHtR, BMI, LDL/HDL, and TC/HDL subsequently. Therefore, LAP and CMI provide a more accurate indication of MetS.

This study also analyzed the cutoff points of LAP and CMI for distinguishing MetS. The identified cutoff points for LAP were found to be 49.87 for women and 52.76 for men. Additionally, the cutoff points for CMI distinguishing MetS were determined to be 0.56 for women and 0.70 for men. The identification of cutoff points facilitates the clinical screening of high-risk MetS populations. It is noteworthy that the optimal thresholds for LAP and CMI in distinguishing MetS may vary depending on the study population’s race, age, and the diagnostic criteria for MetS. Similar to our findings, another study based on a U.S. population identified optimal LAP cutoffs of 52.43 for women and 53.31 for men [29]. However, the Iranian [45], Spanish [46] and China [39] populations had slightly lower optimal thresholds. Comparable cutoff points of CMI were identified in individuals from South India, with all subjects having a value of 0.64 [47]. Conversely, higher cutoff points were observed in individuals with obesity from Italy, where females had a threshold of 1.14 and males had a threshold of 1.47 [48]. Currently, most existing predictive markers are either obesity-related indicators (such as LAP, CMI) or insulin resistance-related markers (such as TyG). Kyung’s findings indicated that LAP’s ability to distinguish MetS was superior to that of Tyg [49], and another study had the same finding [50]. Overall, existing research supports the reliability of LAP and CMI as predictive markers for MetS. Furthermore, LAP and CMI are easy to calculate, cost-effective, and convenient to measure, making them broadly applicable in clinical settings.

Previous research has identified gender disparities in the predictive capacity of LAP and CMI for various metabolic conditions such as diabetes, hypertension, and NAFLD [29, 36, 51]. These differences were commonly attributed to anatomical and physiological distinctions between males and females. Males typically accumulate fat in the abdominal region due to androgenic influences like testosterone, whereas females tend to store fat in the buttocks and thighs due to estrogenic effects [52]. This variance in fat distribution is believed to be the primary factor contributing to sex-based prevalence differences. Studies have shown that the prevalence of MetS in menstruating females is lower than that in males, while the prevalence in females who are peri- or postmenopausal is equivalent to that of males [53]. Although OR values for LAP and CMI in relation to MetS risk were higher in males, further analysis indicated no significant difference in the AUC for distinguishing MetS between the sexes. Given that a substantial proportion of female participants in this study were peri- or postmenopausal, during which decreased estrogen levels result in a shift in fat storage from subcutaneous to visceral fat [52], it is important to note that reduced estrogen also contributes to increased components of MetS, such as elevated blood pressure, blood glucose, and blood lipid levels [54–56]. This shift significantly heightens the risk of MetS in females. Currently, evidence remains inconclusive regarding whether LAP or CMI offers superior predictive value for MetS in males or females. Additionally, our study found that LAP and CMI were associated with all-cause mortality in females but not in males. This discrepancy may be attributed to the protective cardiovascular effects of estrogen in premenopausal women, which diminish after menopause, leading to increased visceral fat accumulation, exacerbated insulin resistance, and a marked rise in the risk of cancer mortality, cardiovascular mortality, and all-cause mortality [56–58]. Supporting this, Cerhan et al. observed in a cohort of 650,000 adults that abdominal fat accumulation was significantly associated with higher all-cause mortality and cardiovascular disease risk in females compared to males [59]. However, our study found that LAP and CMI were not associated with cardiovascular mortality in females, which might result from incomplete adjustment for confounding factors and the heterogeneity of the population.

The precise mechanisms connecting LAP and CMI to MetS remain elusive. It is widely acknowledged that obesity and IR might mediate this relationship. Initially, visceral fat cells release elevated levels of free fatty acids (FFA), which infiltrate liver and muscle tissues, disrupting insulin pathways and leading to IR [60]. Additionally, visceral fat cells produce pro-inflammatory cytokines (such as IL-6, TNF-α) that can also disrupt insulin signaling [61]. Moreover, the infiltration of macrophages and other immune cells in VAT induces a persistent low-grade inflammatory state, which affects insulin sensitivity and lipid metabolism [61]. Furthermore, FFA from VAT find their way to the liver, stimulating the synthesis of very low-density lipoproteins, escalating plasma triglyceride concentrations, reducing HDL-C levels, and increasing the likelihood of atherosclerosis [61]. VAT functions as an endocrine organ, secreting hormones like leptin and adiponectin. Disruptions in these hormonal levels can affect metabolic processes and heighten the risk of MetS [11]. As a result, LAP and CMI appear to be reliable indicators of VAT and IR, acting as straightforward and financially accessible biological tools for distinguishing MetS.

Study strengths and limitations

This research offers various benefits. Initially, it utilized the American NHANES database, providing a substantial sample size that mirrors the entire United States population. Next, it conducted a thorough comparison of new obesity metrics with traditional ones in recognizing MetS. Thirdly, it explored gender-specific differences, shedding light on how gender impacts MetS prediction accuracy. Furthermore, it delved into the relationships between LAP and CMI and all-cause mortality and cardiovascular mortality, offering valuable long-term prognosis insights. Ultimately, the research employed multiple statistical approaches for validation, reinforcing the credibility of the findings. Nevertheless, some limitations were noted. The study design, despite attempts at confounding correction, could not control all potential confounders. Secondly, to decrease recall bias, the focus was mainly on objective measurements and biochemical markers, neglecting subjective recall data on participants’ lifestyle habits. Thirdly, the absence of questionnaire data on lipid-lowering medication use hindered consideration of this aspect. Despite utilizing the American NHANES database, which represents the US population, generalizing the findings to other ethnicities, nations, or regions requires further validation.

Conclusion

In comparison to some analyzed indicators of obesity, LAP and CMI demonstrated superior diagnostic accuracy for MetS. These indicators are affordable, easily accessible, and widely applicable for early identification and screening of MetS in medical settings. Additionally, LAP and CMI levels are linked to the probability of mortality from all causes in both the general population and the female population. This correlation suggests that utilizing LAP and CMI as screening tools for MetS could significantly reduce overall mortality rates in the future.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

MetS

metabolic syndrome

LAP

lipid accumulation products

CMI

cardiometabolic index

ROC

receiver operating characteristic

IR

insulin resistance

CVD

Cardiovascular disease

CT

computed tomography

MRI

magnetic resonance imaging

BMI

body mass index

WC

waist circumference

WHtR

waist-to-height ratio

Tyg

triglyceride-glucose

NAFLD

non-alcoholic fatty liver disease

NHANES

National Health and Nutrition Examination Survey

CDC

Centers for Disease Control and Prevention

TG

triglyceride

FBG

fasting blood glucose

TC

total cholesterol

HDL-C

high-density lipoprotein cholesterol

LDL-C

low-density lipoprotein cholesterol

AST

aspartate aminotransferase

ALT

alanine aminotransferase

HbA1C

glycated hemoglobin A1C

SBP

systolic blood pressure

DBP

diastolic blood pressure

OGTT

oral glucose tolerance test

SD

standard deviation

RCS

Restricted cubic spline

OR

odds ratio

HR

hazard ratios

CI

confidence intervals

AUC

area under the curve

Author contributions

YT. is the project leader, while YT., YHJ., and JHS. collaborated on project design and manuscript review. XJC. and YFZ. conducted data analysis and contributed to manuscript writing. All authors participated in writing and approving the final manuscript.

Funding

No Fundings.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.