This large-scale multicenter retrospective study population comprised 214,717 CVD cases hospitalized in Tianjin between January 1, 2014, and June 30, 2022. The following cases were excluded: (1) those have extreme values; and (2) those who lacked data on Body Mass Index (BMI), TG, High Density Lipoprotein Cholesterol (HDL-C), Low Density Lipoprotein Cholesterol (LDL-C), and Fasting Plasma Glucose (FPG). Based on previous research, we estimated the overall probability as P≈68%, with an allowable error of δ≈0.15P, leading to a calculated sample size of 41 cases (9). To strengthen the robustness of our findings, we maximized the sample size. Ultimately, a total of 17,632 cases were included in the final analysis. The schematic representation illustrating the cases recruitment process is depicted in Figure 1 . Approval for the study was obtained from the Ethics Committee of Tianjin University of Traditional Chinese Medicine (TJUTCM-EC20210007) and registered with the China Clinical Trials Registry (ChiCTR2200058296) and with Clinical Trials. gov (NCT05309343).

Age, sex, smoking, drinking, and medication history of cases were collected through standardized questionnaires, which were administered approximately four hours following their admission (10, 11). Systolic blood pressure (SBP), diastolic blood pressure (DBP), FPG, glycated hemoglobin (HbA1c), TC, TG, HDL-C, and LDL-C levels were recorded by practitioners using standardized laboratory methods (12).

METS-IR was calculated as (ln [(2 × FPG) + TG)] × BMI)/(ln [HDL-C]) (FPG, TG and HDL-C levels were expressed as mg/dl and BMI as kg/m² in the equation) (13). BMI was calculated by dividing weight by square height and was expressed in kg/m². Hyper-triglyceridemia (hyper-TG), hyper-cholestero lemia (hyper-TC), hyper-high-density lipoprotein cholesterolemia (hyper-HDL), hyper-low-density lipoprotein cholesterolemia (hyper-LDL) were defined as TG ≥ 1.7 mmol/L, TC ≥ 5.2 mmol/L, HDL-C < 1.0 mmol/L, and LDL-C ≥ 3.4 mmol/L, respectively, and dyslipidemia if any one of them (14). According to the American Diabetes Association’s Standards for the Medical Treatment of Diabetes (15), diabetes mellitus (DM) is defined as FPG ≥7.0 mmol/L or HbA1c ≥6.5%, prediabetes mellitus (Pre-DM) is defined as 5.6 mmol/L≤FPG ≤ 6.9 mmol/L or 5.7%≤HbA1c ≤ 6.4%, and normal glucose regulation (NGR) is defined as FPG<5.6 mmol/L or HbA1c<5.7%.

Continuous data are presented as mean ± standard deviation (SD) or median and interquartile range (IQR). Categorical variables are expressed as percentages. Differences between the groups were calculated using the χ² test for categorical variables and the t-test Mann-Whitney test or Kruskal-Wallis test for continuous variables. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using logistic regression analysis to test the association between METS-IR index and the risk of dyslipidemia in cases with CVD. To ascertain the precision of the METS-IR index in identifying dyslipidemia among cases diagnosed with CVD, the area under the curve (AUC) was derived from the receiver operating characteristic (ROC) curve analysis. The statistical significance threshold was set at a P-value of less than 0.05. All statistical analyses were performed using the SPSS 24.0 (IBM Corp, New York, NY, USA).

Among the 17,632 cases incorporated into this study, the proportion of males (55.75%) was slightly higher than that of females (44.2%); the median age of cases stood at 65 years, with 11,934 cases exhibiting dyslipidemia. Relative to the normolipidemic cohort, the dyslipidemic group predominantly comprised males, exhibited a lower mean age, registered higher METS-IR scores, and demonstrated a predisposition towards deleterious habits, including smoking and drinking (P < 0.01) ( Table 1 ).

In the context of binary logistic regression analyses, continuous variables revealed a significant and positive association between the METS-IR index and the risk of developing dyslipidemia. Furthermore, the METS-IR index exhibited a positive association with an elevated OR for dyslipidemia in quartiles Q2, Q3, and Q4, with Q1 serving as the reference group. This association persisted as statistically significant even after adjusting for confounding variables. With ascending levels of the METS-IR index, the ORs for dyslipidemia were observed to be 2.34 (95% CI 2.15-2.55), 4.45 (95% CI 4.05-4.89), and 11.94 (95% CI 10.60-13.45), sequentially (all P < 0.001) ( Table 2 ).

Stratified analyses were performed in Table 2 , which were conducted with consideration for the disparities among distinct subcategories of dyslipidemia. The METS-IR index exhibited a positive association with the risk of hyper-TG and hyper-LDL, albeit without significant association concerning the concurrent risks of hyper-TG and hyper-TC. The ORs corresponding to the METS-IR index in conjunction with hyper-TG and hyper-TC for quartiles two through four (Q2, Q3, and Q4), respectively, exhibited a persistent upward trend as the METS-IR value increased, with Q1 serving as the baseline comparator. Importantly, these observed associations sustained their statistical significance following the application of adjustments for a range of confounding variables (all P < 0.001).

Figure 2 illustrates the ROC curves representing dyslipidemia and the METS-IR index among individuals diagnosed with CVD. The area under the curve (AUC) metric for the METS-IR index was calculated to be 0.747 (95% CI 0.739-0.754; P < 0.001), suggesting a statistically significant capacity of the METS-IR index in identifying and diagnosing dyslipidemic conditions.

The employment of binary logistic regression analysis served to evaluate the impact of varying glucose metabolic conditions on the association between the METS-IR index and the presence of dyslipidemia, with findings summarized in Table 3 . Concerning the METS-IR index as a continuously measured variable, a statistically significant association was observed with the incidence of dyslipidemia across all distinct categories of glucose metabolic conditions (P < 0.001). Furthermore, in the context where Q1 served as the reference, the ORs associated with the presence of dyslipidemia were observed to escalate in conjunction with ascending levels of the METS-IR index, with Q4 having the highest OR (all P < 0.01). These associations persisted in their significance following the adjustment of the model parameters.

The current study has several limitations. First, although we accounted for numerous potential confounding factors, dyslipidemia may still be influenced by other variables, such as systemic lupus erythematosus, hereditary dyslipidemia, or cardiac catheterization. Second, this study was conducted in a Chinese population, where racial differences may exist. Third, potential Neyman bias should be considered. Therefore, well-designed randomized controlled trials are needed to validate these results.