Remember me
Aims:
To characterize the metabolic profiles associated with the Hepatic Insulin Resistance Index (HIRI), in Korean patients with type 2 diabetes mellitus.
Methods:
This cross-sectional study analyzed 2,475 Korean adults with type 2 diabetes from the Korean National Diabetes Program cohort. Participants were categorized into HIRI tertiles based on oral glucose tolerance tests. We examined associations between HIRI and anthropometric measures, glycemic parameters, lipid profiles, liver function markers, and dietary intake. Logistic regression assessed the risk of dyslipidemia and metabolic syndrome components, adjusting for relevant confounders.
Results:
High-HIRI participants demonstrated greater central obesity (waist circumference 90.92 vs. 85.52 cm, p < 0.0001) and distinctive glycemic profiles: elevated fasting glucose (153.26 vs. 147.53 mg/dL, p < 0.0001) but lower postprandial glucose (273.73 vs. 299.32 mg/dL, p = 0.0001) and HbA1c (7.8 vs. 8.2%, p = 0.0026). They exhibited dyslipidemia with higher triglycerides (190.71 vs. 141.41 mg/dL, p < 0.0001) and lower HDL cholesterol (45.29 vs. 48.40 mg/dL, p < 0.0001).
Conclusions:
High hepatic insulin resistance in Korean patients with type 2 diabetes defines a phenotype characterized by central obesity, atherogenic dyslipidemia, and distinctive glycemic profiles. Recognition of this phenotype supports personalized diabetes management targeting hepatic IR.
1 IntroductionInsulin resistance (IR) is a core pathophysiological mechanism in type 2 diabetes (T2D) and a key independent risk factor for cardiovascular disease (1, 2). Recent studies have demonstrated that while IR is a systemic phenomenon, it exhibits distinct metabolic characteristics in major target organs including the liver (hepatic), skeletal muscle, and adipose tissue (3).
The relative contribution of IR in each organ results in distinct clinical manifestations (4). For instance, skeletal muscle IR predominantly impairs postprandial glucose disposal, whereas hepatic IR is a primary determinant of fasting hyperglycemia and directly promotes atherogenic dyslipidemia (5). This distinction is clinically crucial because it suggests that a patient’s specific metabolic risk profile—whether dominated by hyperglycemia or dyslipidemia—may depend on the primary site of IR (6). This provides a strong rationale for investigating organ-specific IR beyond systemic measures.
Among these organ-specific resistances, hepatic IR is distinguished from that in skeletal muscle and adipose tissue by its unique pathophysiological characteristics (4). A key phenomenon observed in the liver is “selective hepatic insulin resistance,” a paradoxical state where insulin fails to suppress gluconeogenesis, yet its signaling for de novo lipogenesis is preserved or even enhanced (7). In contrast, skeletal muscle IR is primarily characterized by impaired glucose uptake, while adipose tissue IR manifests as a failure to suppress lipolysis. Therefore, hepatic IR, with its dual impact on glucose and lipid metabolism, simultaneously drives hyperglycemia and dyslipidemia through VLDL overproduction (5).
This dual metabolic burden is particularly pronounced in East Asian populations, who are predisposed to ectopic fat accumulation even at a lower body mass index (8). Consequently, hepatic IR, primarily driven by Metabolic dysfunction-Associated Steatotic Liver Disease (MASLD), represents a key metabolic phenotype in this demographic (9). Furthermore, hepatic fat accumulation is not merely a localized issue; it is a central driver of systemic IR, with strong links to skeletal muscle IR (10).
Therefore, this study aimed to characterize the distinct glucose and lipid metabolic profiles associated with hepatic IR, and to explore their association with cardiovascular risk markers, in Korean patients with T2D. Through this characterization, we sought to provide deeper insights into the underlying pathophysiology and help identify distinct metabolic phenotypes for targeted therapeutic strategies.
2 Methods2.1 Data sourceThis cross-sectional study analyzed data from the Korea National Diabetes Program (KNDP) cohort study. KNDP is a prospective, multicenter observational study initiated in 2006 to enhance clinical and pathophysiological understanding of Korean patients with type 2 diabetes and individuals at high risk for diabetes. The detailed design of the KNDP cohort has been reported previously (11). The observation period was defined from baseline to March 2014. Eligible participants were patients with T2D aged over 19 years. The exclusion criteria were as follows: patients who had no oral glucose tolerance test (OGTT) records and those who could not be analyzed because of data entry errors. A total of 2,475 patients remained for analysis after exclusion criteria were implemented. Individual informed consent was exempted due to the retrospective nature of the data collection. The study protocol was approved by the institutional review boards of the Kyung Hee University Hospital (KMC IRB 0526-04).
2.2 Clinical and laboratory measurementsClinical and laboratory data were extracted from KNDP at the time of patient visits. This included data on diabetes medication use, smoking status, and alcohol consumption. Smoking status was classified as never-smoker (no history of smoking), ever-smoker (≥10 packs of cigarettes over a lifetime), or current smoker (ever-smoker who currently smokes). Alcohol consumption status was classified as never-drinker (no history of drinking), ever-drinker (history of consuming ≥1–2 glasses of beer), or current drinker (ever-drinker who currently consumes alcohol). Physical measurements encompassed height, weight, body mass index (BMI), systolic blood pressure (SBP), and diastolic blood pressure (DBP). BMI was determined using the standard calculation: weight (kg) divided by height squared (m²).
Glucose metabolism markers were obtained through an OGTT. The glucose metabolism parameters analyzed included HbA1c, fasting plasma glucose, and 2-hour postprandial blood glucose. To assess pancreatic beta-cell function and whole-body IR, the Insulinogenic Index (IGI) and the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) were calculated. The IGI was computed as the ratio of the incremental change in insulin to glucose during the first 30 minutes of the OGTT: (ΔInsulin [0–30 min]/ΔGlucose [0–30 min]) (12). HOMA-IR was calculated using the formula: [fasting insulin (μU/mL) × fasting glucose (mmol/L)]/22.5 (13). Additional laboratory parameters encompassed the lipid profile (total cholesterol, triglycerides, LDL cholesterol, and HDL cholesterol), hepatic function tests (AST, ALT, and gamma-glutamyl transferase), and kidney function indicators (blood urea nitrogen, serum creatinine, estimated creatinine clearance, and urine microalbumin). Estimated creatinine clearance was computed using the MDRD equation: 186 × creatinine(-1.154) × age(-0.203) (× 0.742 for females) (14).
2.3 Dietary assessmentDietary intake was assessed using 3-day food records collected from KNDP during their hospital visits to evaluate participants’ usual dietary patterns. The assessment included intake of energy, macronutrients (carbohydrates, protein, and fat), and minerals such as calcium. Dietary analysis was performed using CAN-Pro 5.0 (Korean Nutrition Society, Republic of Korea).
2.4 Definition of dyslipidemia and cardiometabolic risk factorsDyslipidemia was defined as the presence of one or more of the following: high total cholesterol (≥180 mg/dL), high triglycerides (≥150 mg/dL), low HDL cholesterol (<40 mg/dL for men, <50 mg/dL for women), or high LDL cholesterol (≥100 mg/dL), or use of lipid-lowering medication (15). And components of metabolic syndrome were defined based on the Harmonized National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria (16). However, the waist circumference (WC) criteria were adapted for Korean adults, defining high WC as ≥ 90 cm for men and ≥ 85 cm for women (17). Hypertension was defined as a systolic blood pressure ≥130 mmHg, a diastolic blood pressure ≥85 mmHg, or use of antihypertensive medication. Metabolic syndrome was diagnosed if three or more of these risk factors were present.
2.5 Hepatic insulin resistance indexThe hepatic-specific IR index was derived from the OGTT. The Hepatic Insulin Resistance Index (HIRI), which has been validated against the gold standard insulin clamp technique, was calculated by multiplying the area under the curve (AUC) for plasma glucose and plasma insulin concentrations from 0 to 30 minutes during the OGTT (18). This index has been utilized in previous studies and has demonstrated its clinical utility in assessing hepatic insulin sensitivity (19–21). To account for sex differences in hepatic IR, participants were first stratified by sex, and then categorized into tertiles based on their HIRI values within each sex group, following the approach used in previous research (20, 22).
2.6 Statistical analysisAll variables were expressed as mean ± standard deviation for continuous variables or as frequency (%) for categorical variables. Normality testing was performed for all continuous variables, and group differences were assessed using analysis of variance (ANOVA) or Kruskal-Wallis test depending on the distribution of data. When significant differences were identified, post-hoc analysis was conducted using Bonferroni correction for multiple comparisons. Spearman’s rank correlation analysis was used to assess the associations between variables. Logistic regression analysis was performed to examine the incidence of dyslipidemia, metabolic syndrome, and its individual components. The regression models were adjusted for covariates including sex, age, BMI, duration of diabetes, smoking status, and alcohol consumption. A p-value of less than 0.05 was considered statistically significant. All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
3 Results3.1 Study population characteristicsA total of 2,475 Korean patients with type 2 diabetes mellitus were included in the analysis, comprising 1,393 (56%) men and 1,082 (44%) women. Participants were divided into tertiles based on HIRI values: low-HIRI (n = 824), middle-HIRI (n = 826), and high-HIRI (n = 825) groups. The median HIRI values were 113.38 (IQR: 77.58-140.63), 222.16 (IQR: 192.32-258.74), and 435.17 (IQR: 358.02-574.32) for the low, middle, and high tertiles, respectively (Table 1) (Figure 1).
Variables1)Low-HIRI (n = 824)Middle-HIRI (n = 826)High-HIRI (n = 825)P-value†P for trend††Age (year)52.91 ± 10.4953.62 ± 10.3153.40 ± 10.570.35570.2806Height (cm)162.34 ± 8.79162.87 ± 8.58162.85 ± 9.130.49000.3220Weight (kg)63.41 ± 9.97c66.77 ± 10.87b70.74 ± 11.82<.0001<.0001BMI (kg/m2)24.00 ± 2.90c25.10 ± 3.07b26.58 ± 3.25a<.0001<.0001Hip circumference (cm)94.54 ± 6.43c96.35 ± 6.55b98.18 ± 6.57a<.0001<.0001Waist circumference (cm)85.52 ± 7.87c87.94 ± 7.96b90.92 ± 8.27a<.0001<.0001Systolic blood pressure (mmHg)124.95 ± 15.23125.93 ± 15.35126.70 ± 15.130.05570.0165Diastolic blood pressure (mmHg)78.65 ± 9.51b79.29 ± 9.94ab80.75 ± 9.83a0.02320.0062Smoking Never-smokers431 (54)433 (53)432 (53) Ex-smokers187 (23)211 (21)216 (27)0.43950.6327 Current smokers187 (23)168 (26)167 (21)Alcohol consumption status Never-drinkers327 (41)348 (43)354 (43) Ex-drinkers108 (13)92 (11)90 (11)0.53780.2586 Current drinkers369 (46)372 (46)372 (46) Diabetes duration (year)4.87 ± 6.04a4.80 ± 6.19a3.58 ± 4.94b0.00040.0002Anthropometric, lifestyle, and clinical profiles by hepatic insulin resistance index (HIRI) tertiles in Korean type 2 diabetes.
BMI, body mass index (Weight(kg)/Height(m)2).
1)Values are expressed as Mean ± Standard Deviation as continuous variables and Number (%) as categorial variables.
†P-values calculated using ANOVA or Kruskal–Wallis test, depending on data normality.
††P for trends were obtained by the Jonckheere-Terpstra test for continuous variables, Cochran-Mantel-Haenszel statistics for categorical variables.
Values with different superscript letters are significantly different (p < 0.05) by Bonferroni post hoc test.
Bold-faced p-values indicate statistical significance (p-value < 0.05).
Tertiles of the Hepatic Insulin Resistance Index (HIRI) in Korean adults with type 2 diabetes. HIRI was calculated as glucose AUC0–30 × insulin AUC0–30 during the OGTT. Participants were categorized into HIRI tertiles defined separately in men and women and then combined into T1–T3 groups (T1, lowest; T3, highest). Values are presented as median (interquartile range). The horizontal color gradient (green to red) indicates increasing HIRI.
Participants with higher HIRI demonstrated significantly elevated anthropometric measurements compared to those with lower HIRI. While height showed no significant difference across tertiles (p = 0.4900), body weight increased from the low-HIRI to high-HIRI (63.41 ± 9.97 kg vs. 70.74 ± 11.82 kg, p < 0.0001), resulting in correspondingly higher BMI (24.00 ± 2.90 kg/m² vs. 26.58 ± 3.25 kg/m², p < 0.0001). Both hip (94.54 ± 6.43 cm vs. 98.18 ± 6.57 cm, p < 0.0001) and WC (85.52 ± 7.87 cm vs. 90.92 ± 8.27 cm, p < 0.0001) showed positive correlations with HIRI. Additionally, both systolic and diastolic blood pressure exhibited increasing trends across tertiles, with diastolic blood pressure showing statistical significance (78.65 ± 9.51 mmHg vs. 80.75 ± 9.83 mmHg, p = 0.0232). Notably, diabetes duration was significantly shorter in the high-HIRI group compared to the low-HIRI group (3.58 ± 4.94 years vs. 4.87 ± 6.04 years, p = 0.0004), showing a decreasing trend across tertiles (p for trend=0.0002).
3.2 Correlations between HIRI and clinical parametersThe correlations between HIRI and various clinical parameters are summarized in Figure 2. HIRI showed positive correlations with anthropometric measures including BMI and WC. For glycemic parameters, HIRI demonstrated a positive correlation with fasting glucose while showing negative correlations with 2-hour postprandial glucose and HbA1c. Regarding lipid profiles, HIRI was positively correlated with total cholesterol and triglycerides and negatively correlated with HDL cholesterol. HIRI also showed positive correlations with liver function markers, including ALT, AST, and GGT. By contrast, correlations with dietary intake were generally weak or non-significant, although modest associations were observed for certain fatty acid subtypes and carbohydrate intake.
Spearman correlation map between the Hepatic Insulin Resistance Index (HIRI) and metabolic variables. Variables are grouped by major domains (inner ring): age and BMI, anthropometric measures, clinical profiles, glucose parameters, insulin resistance (IR) indices, blood lipids, liver function, kidney function, and dietary intake. Each segment displays Spearman’s correlation coefficient (ρ) between HIRI and the corresponding variable, and the color scale denotes the direction and magnitude of the correlation (blue, negative; red, positive; range −0.3 to 0.3). Statistical significance is indicated by asterisks (*p < 0.05, **p < 0.001, ***p < 0.0001).
3.3 Metabolic profiles associated with HIRIBeyond correlation analysis, we compared metabolic characteristics across HIRI tertiles (Table 2). The high-HIRI group had significantly higher fasting glucose levels compared to the low-HIRI group (153.26 ± 52.49 mg/dL vs. 147.53 ± 59.65 mg/dL), with significant differences observed among the three groups (p < 0.0001). Interestingly, both 2-hour postprandial glucose (273.73 ± 99.55 mg/dL vs. 299.32 ± 110.09 mg/dL; p = 0.0001) and HbA1c (7.79 ± 1.88% vs. 8.18 ± 2.18%; p = 0.0026) were significantly lower in the high-HIRI group. Indices of IR and secretion were markedly higher with increasing HIRI, including fasting insulin, HOMA-IR, and IGI (p < 0.0001). HOMA-IR, for example, was approximately three times higher in the high-HIRI than in the low-HIRI (5.45 ± 4.94 vs. 1.37 ± 0.81).
Variables1)Low-HIRIHepatic insulin resistance index (HIRI) and associated metabolic profiles in Korean patients with type 2 diabetes.
HbA1c: Hemoglobin A1c; HOMA-IR: Homeostatic Model Assessment for Insulin Resistance; HDL: High-Density Lipoprotein; LDL: Low-Density lipoprotein; AST: Aspartate Transaminase; ALT: Alanine Transaminase; BUN: Blood Urea Nitrogen.
1)Values are expressed as Mean ± Standard Deviation as continuous variables.
†P-values calculated using ANOVA or Kruskal–Wallis test, depending on data normality.
Values with different superscript letters are significantly different (p < 0.05) by Bonferroni post hoc test.
Bold-faced p-values indicate statistical significance (p-value < 0.05).
The lipid profile also showed significant associations with HIRI. The high-HIRI demonstrated significantly higher levels of total cholesterol (191.63 ± 41.39 mg/dL vs. 184.11 ± 39.24 mg/dL; p = 0.0018) and triglycerides (190.71 ± 138.06 mg/dL vs. 141.41 ± 90.80 mg/dL; p < 0.0001). In contrast, HDL cholesterol levels were significantly lower in the high-HIRI (45.29 ± 11.63 mg/dL vs. 48.40 ± 12.61 mg/dL; p < 0.0001). There was no significant difference in LDL cholesterol levels among the three groups (p = 0.4244). Liver function markers, including AST, ALT, and Gamma-glutamyl Transferase, were all significantly higher in the high-HIRI compared to the low-HIRI (p < 0.0001).
Regarding kidney function, BUN levels did not differ significantly among the groups (p = 0.8992). However, both serum creatinine (0.83 ± 0.24 mg/dL vs. 0.78 ± 0.22 mg/dL; p < 0.0001) and creatinine clearance (102.71 ± 35.58 mL/min vs. 97.62 ± 31.11 mL/min; p = 0.0023) were significantly higher in the high-HIRI compared to the low-HIRI.
3.4 Dietary nutrient intakeDietary analyses were conducted using available dietary data across the HIRI tertiles (Table 3). There were no significant differences in total daily energy intake (p = 0.3322), total carbohydrate intake (p = 0.4190), or total protein intake (p = 0.1367) among the groups. However, when analyzed as a proportion of total energy or relative to body weight, several differences emerged.
Variables1)Low-HIRIDietary intake according to hepatic insulin resistance index (HIRI) tertile groups in Korean patients with type 2 diabetes.
1)Values are expressed as Mean ± Standard Deviation as continuous variables.
†P-values calculated using ANOVA or Kruskal–Wallis test, depending on data normality.
Values with different superscript letters are significantly different (p < 0.05) by Bonferroni post hoc test.
Bold-faced p-values indicate statistical significance (p-value < 0.05).
The percentage of energy from carbohydrates was significantly lower in the high-HIRI compared to the low-HIRI (59.54 ± 9.57% vs. 61.01 ± 9.52%; p = 0.0078). In contrast, the percentage of energy from fat was significantly higher (22.49 ± 7.41% vs. 21.53 ± 7.98%; p = 0.0062). When adjusted for body weight, protein intake (1.12 ± 0.35 g/kg BW vs. 1.25 ± 0.41 g/kg BW; p < 0.0001) was significantly lower in the high-HIRI.
An analysis of fat composition revealed that the high-HIRI group consumed a significantly greater amount of total fat (46.31 ± 21.11 g vs. 43.97 ± 23.19 g; p = 0.0117). This was reflected across various fatty acid types, with higher intake of total fatty acids (p = 0.0008), saturated fatty acids (p = 0.0022), monounsaturated fatty acids (p = 0.0019), and polyunsaturated fatty acids (p = 0.0017) in the high-HIRI. Intake of dietary fiber, and cholesterol did not differ significantly across the tertiles.
Comments (0)