Diagnostic Performance of Surrogate Indices of Insulin Resistance Compared with HOMA2-IR in Non-Diabetic Patients with Acute Coronary Syndrome

Abstract

Background:

Insulin resistance (IR) plays an important role in metabolic dysregulation and adverse outcomes in acute coronary syndrome (ACS). However, assessment of IR during the acute phase of ACS is challenging, as stress-induced hyperglycemia may confound glucose-based surrogate indices.

Objective:

This study aimed to evaluate the correlation and diagnostic performance of multiple surrogate indices of IR in comparison with HOMA2-IR in non-diabetic patients with ACS during the acute phase of hospitalization.

Methods:

A cross-sectional study was conducted in 116 non-diabetic patients with ACS and 120 healthy control individuals. The healthy control group was used to establish a population-specific cut-off for HOMA2-IR (≥ 1.35). Surrogate indices of IR were calculated from fasting biochemical measurements obtained within 24–48 hours of admission. Diagnostic performance was assessed using the area under the receiver operating characteristic curve (AUC), Spearman’s correlation analysis, multivariable logistic regression, and reclassification metrics, including net reclassification improvement (NRI) and integrated discrimination improvement (IDI).

Results:

The McAuley index demonstrated the highest discriminative ability for IR (AUC = 0.863; 95% CI: 0.794–0.923) and the strongest inverse correlation with HOMA2-IR (ρ = –0.798; p < 0.001). In multivariable analyses adjusted for confounders, the McAuley index remained a strong independent predictor of IR (adjusted OR = 0.302; 95% CI: 0.188–0.485; p < 0.001). TyG-BMI showed an independent association only when BMI was excluded from the covariates, whereas TyG, METS-IR, and lipid ratios showed limited or no diagnostic utility. Reclassification analyses confirmed the superior performance of the McAuley index over non-insulin-based indices (the total NRI ranged from 1.042 to 1.230).

Conclusions:

Among non-diabetic patients with acute coronary syndrome during the acute phase, surrogate indices of IR showed variable performance. Glucose-based indices demonstrated limited diagnostic accuracy, whereas the insulin-based McAuley index showed the strongest concordance with HOMA2-IR and the most consistent diagnostic performance among the indices evaluated.

1. BACKGROUNS

Insulin resistance is a critical metabolic abnormality that contributes to the development and progression of coronary artery disease, even in individuals without diabetes. In the setting of ACS, IR may be amplified by acute inflammatory and neurohormonal responses (1, 2). Because most patients with ACS also exhibit underlying dyslipidemia, characterized by elevated triglyceride levels and reduced HDL-C, early identification of IR is clinically important, given its role in exacerbating metabolic dysfunction and cardiovascular risk (3, 4). During acute ACS, stress-induced surges in catecholamines and cortisol can further confound the assessment of IR, particularly when glucose-based surrogate indices are used. Therefore, evaluating IR in non-diabetic patients with ACS is relevant for understanding disease pathophysiology and improving early cardiovascular risk stratification. Although the hyperinsulinemic-euglycemic clamp is considered the reference method for assessing IR (5), it is impractical in routine cardiovascular clinical practice.

Consequently, HOMA2-IR has been widely adopted as a practical surrogate measure of IR in clinical and cardiovascular research. However, HOMA2-IR primarily reflects the glucose-insulin axis and may not fully capture lipid-related metabolic disturbances that are highly relevant in cardiovascular disease (6, 7). Accordingly, several non-insulin-based indices, including the triglyceride-glucose (TyG) index, TyG-BMI, METS-IR (Metabolic Score for IR), and lipid ratios, have been proposed as simple and accessible tools, particularly in cardiovascular settings where insulin measurements are not routinely available (8). In addition, the McAuley index, an insulin-based surrogate incorporating triglyceride levels may better reflect lipotoxicity and disturbances in lipid metabolism that are central to atherosclerotic cardiovascular disease.

2. OBJECTIVE

Given the limited evidence in non-diabetic patients, we examined the correlation and diagnostic performance of multiple surrogate indices of IR in comparison with HOMA2-IR, with the aim of identifying the most appropriate surrogate index for use in the acute cardiovascular setting.

3. MATERIAL AND METHODS

Study design and population

This cross-sectional study included non-diabetic ACS patients who were admitted to the Cardiology and Interventional Cardiology Departments at the University Medical Center Ho Chi Minh City from April 2024 to February 2025. A total of 120 healthy adults were recruited as controls. All control participants were free of metabolic and cardiovascular diseases, were non-obese, and had no clinical or diagnostic evidence of coronary artery pathology. Controls were not age- or sex-matched, as they were used solely to derive a physiological HOMA2-IR cut-off in this cross-sectional study.

Inclusion criteria (ACS group)

Patients were eligible for inclusion in the ACS group if they had a confirmed diagnosis of ACS, including ST-Elevation Myocardial Infarction, non-ST-Elevation Myocardial Infarction, or Unstable Angina according to the 2023 ESC guidelines (9). Patients with known diabetes or admission HbA1c ≥ 6.5% were excluded (10). Eligible patients were required to have available fasting (≥ 8 hours) glucose and insulin measurements obtained within 24–48 hours of hospitalization.

Exclusion criteria

Exclusion criteria included pre-existing type 2 diabetes mellitus or admission HbA1c ≥ 6.5%; severe infection; administration of dextrose-containing intravenous fluids during the fasting window; progressive liver failure; current use of corticosteroids or insulin-sensitizing agents; vasopressor support (e.g., noradrenaline or dobutamine) or emergency corticosteroid therapy; incomplete clinical or biochemical data, severe anemia (hemoglobin < 8 g/dL); known hemoglobinopathies; and refusal to participate.

Sample size calculation

$$\mathrm{n}=3+{\left(\frac{{\mathrm{Z}}_{1-\alpha /2}+{\mathrm{Z}}_{1-\beta }}{0.5\times \ln (\frac{1+\mathrm{r}}{1-\mathrm{r}})}\right)}^2$$

Sample size was based on an expected correlation coefficient of r = 0.30 between surrogate indices and HOMA2-IR, α = 0.05, and power = 85% [11]. A minimum of 97 subjects was required. A total of 116 ACS patients were enrolled.

Data collection

Body mass index (BMI) was calculated, and waist circumference was measured at the umbilical level in the standing position. After an 8-hour overnight fast, venous blood samples were collected from the antecubital vein. Fasting plasma glucose, total cholesterol, triglycerides, HDL-C, LDL-C, and creatinine were analyzed using an automated chemistry analyzer (Olympus AU800, Beckman Coulter, USA). Fasting insulin was measured by chemiluminescent immunoassay (Cobas 8000 e801, Roche, Switzerland). Estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI 2021 equation.

Statistical analysis

Statistical analyses were performed using SPSS version 26.0, R version 4.3.1, and Python version 3.12.1. Continuous variables were summarized as mean ± standard deviation for normally distributed data and as median (interquartile range–IQR) for non-normally distributed data. Between-group comparisons were performed using the independent t-test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables, including surrogate indices of IR. Multivariable logistic regression analysis was used to identify independent predictors of IR.

Optimal cut-offs were determined using the Youden index. Models for the McAuley index and TyG index were adjusted for age, sex, BMI, systolic blood pressure, and eGFR, whereas BMI was excluded from the TyG-BMI and METS-IR models to avoid collinearity. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported. The small unit-based OR reflects the large numerical scale of the TyG-BMI index. Model performance was evaluated using Nagelkerke’s R² and the C-statistic. Continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were calculated based on changes in predicted probabilities derived from logistic regression models. Bootstrap resampling with 500 iterations was performed to estimate 95% confidence intervals and p-values. This number of iterations was considered adequate given the sample size, and the results were stable across resamples. A two-sided p < 0.05 was considered statistically significant.

Assessment of IR

IR was assessed using the Homeostasis Model Assessment 2 (HOMA2-IR), which was calculated with the official HOMA2 Calculator developed by the University of Oxford (available from the University of Oxford). Quartiles of HOMA2-IR were defined based on its distribution in the healthy control group, with the upper quartile (≥ 75^th percentile, HOMA2-IR ≥ 1.35) being used as the population-specific cut-off to define IR in this study.

Surrogate Indices:

• McAuley Index = exp(2.63 – 0.28 × ln[fasting insulin (µU/mL)] – 0.31 × ln[triglyceride (mmol/L]) [12]

• TyG Index = ln (triglyceride (mg/dL) × FPG (mg/dL) / 2) [13]

• TyG-BMI = ln (triglyceride (mg/dL) × FPG (mg/dL) / 2) × BMI (kg/m²) [14]

• METS-IR = ln [2 × FPG (mg/dL) + triglyceride (mg/dL)] × BMI (kg/m²)/ln [HDL-C (mg/dL)] [15]

• TG/HDL-C ratio = triglyceride (mg/dL) / HDL-C (mg/dL) [16]

• TC/HDL-C ratio = total cholesterol (mg/dL) / HDL-C (mg/dL) [17]

Indices were calculated using conventional units according to their original definitions.

Triglyceride levels were converted from mg/dL to mmol/L using the division factor 88.57 prior to application in the McAuley formula.

Institutional Review Board Approval

The study was approved by the Ethics Committee of the University of Medicine and Pharmacy at Ho Chi Minh City (Approval No. 932/HĐĐĐ-ĐHYD). The study was registered at ClinicalTrials.gov (NCT06163768). Ethical Compliance with Human Study: This study was conducted in compliance with the ethical standards of the responsible institutional committee on human experimentation and with the principles of the Declaration of Helsinki.

4. RESULTS

Baseline characteristics of the healthy control population

The healthy control group consisted of 120 adults without known metabolic or cardiovascular diseases. This group was used to characterize the physiological distribution of HOMA2-IR and to derive a population-specific cut-off for IR.

Q: Quartile; Data presented as Mean ± SD or Median (IQR)

In the healthy control population (n = 120), 54 (45%) were male. The median age was 36 years (IQR: 33–41), and the mean BMI was 21.50 ± 4.10 kg/m². the median HOMA2-IR was 0.92 (IQR: 0.65–1.35). The 25^th, 50^th, and 75^th percentile values of HOMA2-IR were 0.65, 0.92, and 1.35, respectively. The 75^th percentile value (HOMA2-IR ≥ 1.35) was used as the population-specific cut-off to define IR in this study.

Baseline characteristics of patients with ACS stratified by IR status

A total of 116 non-diabetic patients with ACS during the acute phase of hospitalization were included. Patients with IR ≥ 1.35 were younger and had higher systolic blood pressure than those without IR (p = 0.027 and p = 0.030, respectively). BMI, fasting insulin, and fasting glucose levels were significantly higher in the IR group (all p < 0.01). No significant differences were observed in eGFR or lipid profiles between the two groups (all p > 0.05).

Baseline IR indices of the study population stratified by IR status

The analysis of surrogate indices across HOMA2-IR categories revealed no significant differences in the TG to HDL-C ratio, the TC to HDL-C ratio, or TyG values. In contrast, TyG-BMI, METS-IR, and the McAuley index showed significant differences between individuals with and without IR (p < 0.001, p = 0.009, and p < 0.001, respectively).

Optimal Cut-off values of surrogate indices for predicting IR

The McAuley index demonstrated the highest diagnostic performance for IR, with high sensitivity (96.65%), moderate specificity (66.67%), and the greatest discriminative ability on receiver operating characteristic analysis (AUC = 0.863; p < 0.001). TyG-BMI and METS-IR showed moderate diagnostic accuracy, whereas TyG and lipid-based ratios exhibited limited discriminatory value and did not reach statistical significance.

Table 1. Baseline characteristics of the healthy control group.

Variables	Control (n = 120)
Male, n (%)	54 (45%)
Age (years)	36 (33–41)
BMI (kg/m²)	21.50 ± 4.10
Fasting Glucose (mg/dL)	90.02 (86.02–96.11)
Fasting Insulin (µU/mL)	6.8 (4.61–10.42)
HOMA2-IR	0.92 (0.65–1.35)
Q1 (≤ 25th percentile)	≤ 0.65
Q2 (26th-50th percentile)	0.66–0.92
Q3 (51st-74th percentile)	0.93–1.34
Q4 (≥ 75th percentile)	≥ 1.35

Table 2. Baseline characteristics of patients with ACS stratified by IR. Data presented as Mean ± SD or Median (IQR). p-values were derived from the independent t-test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables.

Variables	non-IR (n = 48)	IR ≥ 1.35 (n = 68)	p-value
Age (years)	70 ± 12	65 ± 13	0.027
Systolic blood pressure (mmHg)	117.00 (100–126)	121 (113–140)	0.030
BMI (kg/m²)	21.75 ± 2.65	23.68 ± 2.80	< 0.001
eGFR (mL/min/1.73 m²)	76.00 (60.25–87.50)	83.50 (59.75–89.00)	0.482
Fasting Insulin (µU/mL)	6.28 (4.33–7.54)	13.60 (11.23–25.98)	< 0.001
Fasting Glucose (mg/dL)	103.55 (88.75–126.25)	117.00 (103.00–133.88)	0.008
Total cholesterol (mg/dL)	163.75 ± 48.91	163.68 ± 48.62	0.994
Triglycerides (mg/dL)	126.50 (101.50–152.00)	147.00 (95.25–181.75)	0.362
HDL-C (mg/dL)	37.50 (32.00–45.00)	41.50 (36.00–47.00)	0.145
LDL-C (mg/dL)	102.50 (78.75–136.25)	107.50 (79.25–131.50)	0.849

Figure 1. Comparison of surrogate indices between non-IR and IR groups. Boxplots show differences in (a) TG/HDL-C, (b) TC/HDL-C, (c) TyG, (d) TyG-BMI, (e) METS-IR, and (f) McAuley index between participants without IR and with IR, defined by HOMA2-IR. Boxes represent the interquartile range with median lines; whiskers indicate 1.5×IQR and dots represent individual values. Group comparisons were performed using the Mann-Whitney U test, with p -values shown above each panel.

Table 3. Diagnostic performance of surrogate indices for IR. CI: confidence interval; AUC: area under the curve; PPV: Positive predictive value; NPV: Negative predictive value; DeLong’s test was used to test whether AUCs differed from 0.5.

Surrogate Index	AUC (95% CI)	Cut-off	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	p-value
TyG	0.592 (0.479–0.693)	8.90	63.24	60.42	69.35	53.70	0.083
METS-IR	0.643 (0.529–0.744)	34.47	75.00	54.17	69.86	60.47	0.005
TyG-BMI	0.689 (0.601–0.785)	195.31	75.00	58.33	71.83	62.22	< 0.001
McAuley	0.863 (0.794–0.923)	7.19	96.65	66.67	79.75	86.49	< 0.001
TG/HDL-C	0.496 (0.384–0.591)	3.0	64.71	45.83	62.86	48.73	0.942
TC/HDL-C	0.588 (0.481–0.692)	4.21	66.18	54.17	67.16	53.06	0.099

Correlations between HOMA2-IR and surrogate indices of IR

In Spearman correlation analyses, the McAuley index showed the strongest inverse correlation with HOMA2-IR (ρ = −0.798; p < 0.001). TyG-BMI and METS-IR demonstrated moderate positive correlations (ρ = 0.400 and ρ = 0.316; both p < 0.001), while TyG showed a weak but significant correlation (ρ = 0.222; p = 0.016). The TG to HDL-C and TC to HDL-C ratios were not significantly correlated with HOMA2-IR.

Multivariable logistic regression analysis for IR

All models were adjusted for age, sex, systolic blood pressure, and eGFR; BMI was included only for indices that do not incorporate BMI to avoid multicollinearity.

After adjustment for age, sex, body mass index, systolic blood pressure, and estimated glomerular filtration rate, only the McAuley index remained independently associated with IR (adjusted OR = 0.302; 95% CI: 0.188–0.485; p < 0.001) and demonstrated the highest explanatory power (Nagelkerke R² ≈ 0.37). After exclusion of BMI, TyG-BMI also showed an independent association (adjusted OR = 1.019; 95% CI: 1.004–1.035; p = 0.013), whereas METS-IR did not (p = 0.112). When standardized per one standard deviation increase, TyG-BMI demonstrated a substantial association with IR risk (Adjusted OR = 1.82), highlighting that the small unit-based OR (1.019) is due to the large scale of the index.

Table 4. Comparison of multivariable logistic regression models for IR. CI: confidence interval. *OR per 1 SD reflects the effect of a one-standard deviation increase in each index.

Surrogate index (main predictor)	Adjusted OR (95% CI)	p-value	Adjusted OR* (per 1 SD increase)
McAuley	0.302 (0.188–0.485)	< 0.001	0.129
TyG-BMI	1.019 (1.004–1.035)	0.013	1.821
METS-IR	1.064 (0.986–1.149)	0.112	1.428
TyG	1.091 (0.471–2.528)	0.839	1.049

NRI/IDI analysis comparing McAuley with other surrogate indices

Continuous NRI and IDI were calculated from changes in predicted probabilities between the McAuley-based reference model and alternative models. All models were adjusted for age, sex, systolic blood pressure, and eGFR; BMI was included only for indices that do not incorporate BMI to avoid multicollinearity.

When compared with alternative surrogate indices, the McAuley-based model demonstrated a significant improvement in risk reclassification for IR. Replacing TyG, TyG-BMI, METS-IR, and lipid-based ratios with the McAuley index resulted in consistently positive continuous net reclassification improvement values, with total NRI ranging from 1.042 to 1.230. The improvement was observed in both the IR and non-IR groups. In addition, the McAuley model showed significantly higher discrimination ability, with IDI values ranging from 0.257 to 0.320. Bootstrap analyses using 500 resamples confirmed the robustness of these findings, as all total NRI and IDI estimates were statistically significant (p < 0.01).

Figure 2. Correlations between HOMA2-IR and surrogate indices of IR. Scatter plots show the associations of HOMA2-IR with (a) TG/HDL-C, (b) TC/HDL-C, (c) TyG, (d) TyG-BMI, (e) METS-IR, and (f) McAuley index. Solid lines indicate fitted trends for visualization. Correlations were assessed using Spearman’s rank correlation coefficient (ρ), with corresponding two-tailed p-values shown in each panel.

Table 5. NRI and IDI when replacing alternative indices with the McAuley index.

Baseline index	NRI (non-IR group)	NRI (IR group)	Total NRI (95% CI)	p-value	IDI (95% CI)	p-value
TyG	0.583	0.588	1.172 (0.625–1.547)	0.004	0.287 (0.126–0.434)	0.004
TyG-BMI	0.583	0.559	1.142 (0.506–1.533)	0.004	0.289 (0.131–0.468)	0.004
METS-IR	0.583	0.647	1.230 (0.725–1.477)	0.004	0.320 (0.172–0.460)	0.004
TG/HDL-C	0.583	0.500	1.083 (0.578–1.483)	0.004	0.265 (0.082–0.459)	0.008
TC/HDL-C	0.542	0.500	1.042 (0.403–1.412)	0.004	0.257 (0.093–0.440)	0.004
CI: confidence interval. NRI: Net Reclassification Improvement; IDI: Integrated Discrimination Improvement; NRI and IDI were calculated based on predicted probabilities derived from logistic regression models.

5. DISCUSSION

In the healthy control group, the median HOMA2-IR was 0.92, with an interquartile range of 0.65 to 1.35. IR was defined as a HOMA2-IR value of at least 1.35, corresponding to the 75^th percentile in the control population, which was consistent with prior epidemiological studies. This cut-off value was higher than that reported by Aliyu et al. in the Qatar Biobank population and slightly lower than the threshold identified by Kevser et al. in the Turkish population (18). These differences may reflect ethnic-specific characteristics. Asian populations, typically have lower BMI values, as observed in our cohort. However, they are more prone to visceral adiposity, which predisposes them to IR at lower BMI thresholds compared with Western or Middle Eastern populations. Notably, patients with IR were slightly younger and had higher systolic blood pressure. This pattern may partly reflect a selection effect, whereby older patients with ACS and more severe IR may have already developed diabetes and were therefore excluded from the study, resulting in an older remaining cohort with a lower observed prevalence of IR.

In this study, we evaluated surrogate indices of IR in non-diabetic patients during the acute phase of ACS. The McAuley index showed the highest discriminative ability for IR and the strongest inverse correlation with HOMA2-IR. After multivariable adjustment, the McAuley index remained the only independent predictor of IR. When BMI was excluded from the model, TyG-BMI also showed an independent association, whereas METS-IR did not, further underscoring the robustness of the McAuley index. In contrast to population-based studies such as the Qatar Biobank (7), where the TyG index demonstrated the strongest discriminative ability for IR (AUC = 0.92), our study found that the McAuley index was the most robust marker of IR in non-diabetic patients with ACS, exhibiting the highest AUC (0.863) and remaining the only independent predictor after multivariable adjustment. In comparison, TyG and related indices showed limited discriminatory performance in the ACS setting. This limited performance likely reflected acute metabolic disturbances characteristic of ACS, including stress-induced hyperglycemia and dynamic fluctuations in triglyceride levels related to enhanced lipolysis, fasting status, and early statin therapy (19, 20).

In the acute phase of ACS, surges in catecholamines and cortisol stimulate glycogenolysis and gluconeogenesis, leading to transient peripheral IR and stress hyperglycemia. Concurrent activation of alpha-2 adrenergic receptors suppresses pancreatic beta-cell insulin secretion, resulting in a disproportionate increase in plasma glucose relative to circulating insulin (21). Because the TyG index is derived from plasma glucose concentrations, stress hyperglycemia can artificially elevate its value. This may lead to misclassification of underlying IR. Moreover, early administration of high-intensity statins or heparin (prior to fasting sample) can alter triglyceride levels, further confounding the TyG index.

In contrast, the McAuley index incorporates fasting insulin and triglyceride levels and may therefore be less susceptible to stress hyperglycemia-driven distortion than glucose-based indices. Its performance was consistent with findings from the Qatar cohort (AUC = 0.880), in which the McAuley index demonstrated similarly high discriminative ability, supporting its reliability even in the setting of acute cardiovascular disease (7). Supporting evidence from Liu et al. identified TyG-BMI as an independent predictor of contrast-induced acute kidney injury, with an AUC of 0.812 higher discriminative ability than TyG alone (22). In addition, a recent review by Song et al. highlighted TyG-BMI as a sensitive and specific surrogate marker of IR, with reported associations across a broad spectrum of cardiometabolic and cardiovascular outcomes, including coronary artery disease, heart failure, atrial fibrillation, and cardiovascular mortality (23). TyG-BMI may outperform TyG by incorporating BMI, thereby better capturing visceral adiposity, which is an important contributor to IR and is not reflected by TyG alone.

Spearman correlation analysis showed that the McAuley index had a strong inverse correlation with HOMA2-IR, whereas TyG-BMI and METS-IR showed moderate correlations and TyG and lipid-based ratios showed only weak correlations. This finding was consistent with the performance analysis reported by Aliyu et al., in which indices reflecting insulin sensitivity, such as the McAuley index, demonstrated discriminative ability for IR that was comparable to or greater than that of indices based solely on lipid and glucose parameters (7). These observations suggest that indices integrating multiple metabolic components, particularly those incorporating an insulin component, show higher concordance with IR.

Beyond receiver operating characteristic analysis, reclassification analyses demonstrated that the McAuley index improved risk stratification for IR compared with other surrogate indices. Positive and statistically significant total NRI values indicated improved classification in both IR and non-IR groups, while consistently positive IDI values further supported the discriminative performance of the McAuley-based model. These findings remained robust after bootstrap validation, despite the relatively modest sample size. Although TyG-BMI and METS-IR exhibited moderate discriminative ability, consistent with their multicomponent composition, the McAuley index showed the highest overall performance among the evaluated indices (8). In this acute setting, hyperglycemia primarily reflects an acute stress-related neuroendocrine response rather than chronic IR (24). At the same time, early heparin therapy rapidly lowers circulating triglyceride levels through lipoprotein lipase activation, resulting in a transient and artificial reduction in triglyceride concentrations. As a result, the diagnostic performance of glucose-triglyceride-based indices, such as the TyG index, may be attenuated under these acute physiological conditions (25). In contrast, the insulin-based McAuley index is less affected by these rapid changes, thereby retaining its relative prognostic advantage in the early ACS setting (26). From a clinical perspective, these findings suggest that insulin-based indices, particularly the McAuley index, may provide a more reliable assessment of IR under acute cardiovascular stress. Although fasting insulin is not routinely measured in acute cardiovascular care, its inclusion may provide a practical approach for early cardiometabolic risk assessment.

This study has several limitations. HOMA2-IR was used as a pragmatic reference instead of the hyperinsulinemic-euglycemic clamp, which may limit direct inference of true insulin sensitivity. Blood samples were obtained during the acute phase of hospitalization, when early therapeutic interventions and stress-related metabolic changes may have transiently influenced lipid parameters, particularly triglyceride levels, potentially affecting the performance of glucose- and lipid-based indices. In addition, the cross-sectional design precluded evaluation of longitudinal changes in IR and its association with long-term clinical outcomes.

6. CONCLUSION

In non-diabetic patients with acute coronary syndrome during the acute phase, surrogate indices of IR showed variable performance when compared with HOMA2-IR. Glucose-based indices demonstrated limited accuracy, whereas the insulin-based McAuley index, which incorporates triglyceride levels, showed the strongest concordance with HOMA2-IR and the most consistent diagnostic performance among the indices evaluated.

ABBREVIATIONS: ACS - acute coronary syndrome, AUC–area under the curve, BMI–body mass index, ESC - European Society of Cardiology, HDL-C - high-density lipoprotein cholesterol, HOMA2-IR - homeostatic model assessment of insulin resistance, IDI - integrated discrimination improvement, IR - insulin resistance, METS-IR - metabolic score for insulin resistance, NRI - net reclassification improvement, NPV - negative predictive value, OR–odds ratio, PPV - positive predictive value, TC–total cholesterol, TG - triglyceride, TyG - triglyceride–glucose index, TyG-BMI–triglyceride-glucose index with body mass index

Informed Consent:

Written informed consent was obtained from all participants or their legally authorized representatives prior to inclusion in the study.

Data Availability:

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Author Contributions:

Nguyen Ngoc Minh Thu, Lam Vinh Nien and Nguyen Chi Thanh conceived and designed the study. Nguyen Ngoc Minh Thu, Bui The Dung, Nguyen Huu Thinh, and Vu Hoang Vu collected the data. Nguyen Ngoc Minh Thu, Lam Vinh Nien and Nguyen Chi Thanh performed the data analysis. Nguyen Ngoc Minh Thu drafted the manuscript. All authors contributed to interpreting the results, critically reviewed the manuscript, and approved the final version.

Conflict of Interest:

The authors declare that they have no conflicts of interest relevant to this study.

Financial Disclosure or Funding:

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.