Reduced insulin clearance is linked to subclinical atherosclerosis in individuals at risk for type 2 diabetes mellitus

Type 2 diabetes mellitus (T2D) increases risk for atherosclerosis and cardiovascular disease (CVD). Carotid intima-media thickness (cIMT) is a readily accessible surrogate of subclinical atherosclerosis that can detect structural vascular changes years before overt CVD. Multiple mechanisms contribute to increased cIMT, including hyperglycemia and insulin resistance with resulting hyperinsulinemia. Insulin is subject to substantial first-pass hepatic clearance, while C-peptide is not; reduced hepatic insulin clearance elevates systemic insulin levels. Prior work suggests reduced insulin clearance is associated with atherosclerosis independent of insulin-stimulated glucose disposal, but the relative contributions of glycemia, insulin secretion, insulin sensitivity, and insulin clearance to cIMT—beyond classical CVD risk factors—remain unclear. This study aimed to systematically quantify the relative contributions of fasting- and OGTT-derived glycemic and insulin traits, alongside traditional CVD risk factors, to cIMT in individuals at elevated risk for T2D.

Epidemiological and clinical studies have linked elevated fasting, postprandial, and long-term glycemia with increased CVD risk in individuals with diabetes, prediabetes, and even normoglycemia. Insulin resistance and resultant hyperinsulinemia may also contribute to vascular damage. Approximately half of secreted insulin is cleared by the liver during first pass; diminished insulin clearance increases circulating insulin. Prior reports have associated insulin clearance with cIMT and subclinical atherosclerosis, though mechanisms are incompletely understood and confounding by other metabolic risk factors has not been systematically addressed. Insulin signaling exhibits both anti-atherogenic (PI3K-Akt-eNOS) and pro-atherogenic (MAPK-Erk–mediated vascular smooth muscle proliferation) axes, suggesting that chronic hyperinsulinemia from reduced clearance could bias signaling toward pro-atherogenic pathways. The present study builds on this literature by evaluating insulin clearance relative to a broad set of metabolic and traditional CVD risk factors.

Study design and participants: Cross-sectional analysis within the Tübingen Lifestyle Intervention Program of adults at increased risk for T2D. Inclusion required at least one: family history of T2D, BMI > 27 kg/m², prior impaired glucose tolerance, or prior gestational diabetes. Exclusions included known diabetes at baseline and any history of macrovascular CVD. Of 754 screened, 87 were excluded for incomplete OGTT (n = 11) or missing blood pressure (n = 76), leaving 667 participants (417 women, 250 men; mean age 44.1 years). The study was approved by the local ethics committee; all participants provided written informed consent. Assessments: After a 10-hour overnight fast, a 75 g OGTT with samples at 0, 30, 60, and 120 minutes was performed. Measurements included plasma glucose (glucose oxidase method), insulin and C-peptide (ADVIA Centaur XPT), proinsulin (ELISA), HbA1c (Tosoh HLC-723 G8), lipids and hsCRP (ADVIA XPT), and estimated glomerular filtration rate (MDRD formula). Liver fat content was quantified by 1H-MRS on a 1.5 T scanner in a subset (n = 484). Blood pressure was measured after 10 minutes seated rest. Derived indices: Insulin sensitivity was estimated using the Matsuda index. Fasting insulin clearance was calculated as the fasting C-peptide/insulin ratio. Glucose-stimulated (OGTT-derived) insulin clearance was calculated using C-peptide to insulin area-under-the-curve ratios across the OGTT. Additional OGTT-derived indices (e.g., oral disposition index) were calculated per prior publications. cIMT measurement: High-resolution B-mode ultrasound of the common carotid arteries was performed in the early morning prior to OGTT by an experienced examiner using a 10–13 MHz linear transducer (AUS Harmony). Three measurements per side were averaged; the overall mean cIMT was the average of both sides, following European consensus recommendations. Statistical analysis: Variables with skewed distributions were log-transformed. Univariate associations used correlation/regression as appropriate. Multivariate linear regression with standardized beta coefficients was used to identify independent associations with cIMT. Stepwise forward selection with fivefold cross-validation identified robust determinants; results are reported as k-fold r². Significance threshold was p < 0.05. Analyses were conducted in JMP 13.0 (SAS).

- In univariate analyses, classical CVD risk factors (age, waist circumference, blood pressure, dyslipidemia) were associated with higher cIMT; insulin sensitivity was inversely associated with cIMT. Glucose-stimulated insulin clearance was inversely correlated with cIMT (r_s = -0.17, p < 0.0001), whereas fasting insulin clearance was borderline (p ≈ 0.05). - Stepwise forward regression with fivefold cross-validation identified as determinants of cIMT: age (k-fold r² = 0.3936, p < 0.0001), glucose-stimulated insulin clearance (0.4588, p < 0.0001), systolic blood pressure (0.4678, p < 0.0001), BMI (0.4832, p = 0.002), gender (0.4852, p = 0.013), and fasting insulin clearance (0.4854, p = 0.030). Other variables, including HbA1c, hsCRP, and insulin sensitivity, were not retained as significant in this model. - Multiple linear regression (core model) showed independent associations with cIMT: age (β = 0.60, SE 0.02, p < 0.0001), gender (β = -0.08, SE 0.01, p = 0.001), BMI (β = 0.11, SE 0.04, p = 0.001), systolic blood pressure (β = 0.12, SE 0.05, p = 0.05), and glucose-stimulated insulin clearance (β = -0.16, SE 0.02, p < 0.0001). - In an extended model including LDL-cholesterol, hsCRP, and smoking, glucose-stimulated insulin clearance remained independently and inversely associated with cIMT (β = -0.16, p < 0.0001). LDL-cholesterol (p = 0.71), hsCRP (p = 0.06), and smoking (p = 0.98) were not significant. - Subgroup analyses: Univariate associations between insulin clearance and cIMT were significant in newly diagnosed diabetes (β = -0.39, p = 0.03), individuals without diabetes (β = -0.11, p = 0.03), and IFG (β = -0.27, p = 0.01), but not in IGT (p = 0.16) or combined IFG+IGT (p = 0.23). Forward stepwise regression indicated insulin clearance as an independent determinant in newly diagnosed diabetes (r = -0.282, p = 0.037) and in those without diabetes (r = -0.431, p < 0.0001). In newly diagnosed diabetes, age (r² = 0.3995, p < 0.0001) and insulin clearance (r² = 0.4685, p < 0.0001) were the strongest covariates. - No interaction of insulin clearance with gender (p = 0.91) or liver fat content (unadjusted p = 0.72; adjusted p = 0.24). The association of insulin clearance with cIMT persisted after adjusting for liver fat, age, and gender (β = -0.25, p < 0.0001), and liver fat was not selected as a determinant in stepwise analysis (r² = 0.5147, p = 0.53). - Estimated glomerular filtration rate showed no association with cIMT in the full cohort or after excluding impaired kidney function; there was no interaction between eGFR and insulin clearance (p = 0.25).

This study demonstrates that reduced glucose-stimulated insulin clearance is independently associated with greater carotid intima-media thickness in adults at elevated risk for T2D, even after accounting for age, adiposity, blood pressure, and additional CVD risk factors such as LDL-cholesterol, smoking, and hsCRP. The findings suggest that impaired insulin clearance may contribute to early vascular remodeling beyond the effects of hyperglycemia and classical risk factors. Mechanistically, decreased clearance may drive chronic hyperinsulinemia, shifting vascular insulin signaling from PI3K-Akt–mediated vasodilatory and anti-inflammatory effects toward MAPK-Erk–mediated vasoconstriction, smooth muscle proliferation, and inflammation, thereby promoting atherogenesis. The persistence of the association across glycemic categories, including normoglycemia and newly diagnosed diabetes, indicates that the relationship is not solely dependent on ambient glucose metabolism. Lack of interaction with liver fat and kidney function suggests the observed effect is not explained by hepatic steatosis or renal impairment. Overall, the results support considering insulin clearance as a distinct vascular risk determinant and potential early marker of subclinical atherosclerosis.

In individuals at risk for T2D, reduced glucose-stimulated insulin clearance is independently associated with higher cIMT, indicating early subclinical vascular damage beyond traditional CVD risk factors and insulin resistance. Insulin clearance should be considered in risk stratification for cardiovascular disease and as a target for mechanistic and interventional research. Future prospective studies should confirm causality, delineate hepatic versus peripheral contributions to clearance, and evaluate whether improving insulin clearance or mitigating hyperinsulinemia favorably impacts vascular outcomes.

Cross-sectional design precludes causal inference. Insulin clearance was estimated from OGTT-derived measures rather than gold-standard clamp-based assessments. Medication use was not systematically analyzed, leaving potential residual confounding. Some subgroup sample sizes were small, limiting statistical power to detect associations within all glycemic categories. Single-center cohort may limit generalizability.