Influence of Postprandial Hyperglycemic Conditions on Arterial Stiffness in Patients With Type 2 Diabetes

Abstract

Context: Patients with type 2 diabetes (T2D) are at an increased risk of cardiovascular disease.

Objective: The objective of the study was to determine whether postprandial hyperglycemia affects arterial function in T2D.

Design: A single-center, open-label study of three groups of men were studied: 1) T2D patients with albuminuria (n = 22), 2) T2D patients without albuminuria (n = 24), and 3) nondiabetic controls (n = 25). Patients were randomized to a two-period crossover study schedule, ingesting breakfast, with or without insulin lispro (to induce low or high postprandial glycemia).

Main outcome measures: Arterial stiffness was assessed by calculating pulse wave velocity (PWV) and augmentation index using applanation tonometry, and endothelial dysfunction was assessed using peripheral arterial tonometry, 30 minutes before breakfast and up to 240 minutes after breakfast.

Results: At baseline, arterial stiffness was increased in patients. When adjusted for age and body mass index, in a combined group of patients with and without albuminuria, brachial PWV was higher during low (P = .032) and high (P = .038) postprandial glycemia vs controls. These differences were driven by the albuminuria group vs controls during low (P = .014) and high (P = .018) postprandial glycemia. No differences were observed in aortic PWV, augmentation index, or peripheral arterial tonometry ratio between patients and controls. Endothelin-1 and IL-6 were higher, and superoxide dismutase was lower, during postprandial hyperglycemia in T2D patients vs controls.

Conclusions: In patients with T2D and albuminuria, brachial PWV was higher under postprandial hyperglycemic conditions, relative to controls. These data suggest that hyperglycemia induces an increase in stiffness of intermediate-sized arteries. We found no changes in other parts of the arterial bed.

Trial registration: ClinicalTrials.gov NCT01159938.

Figures

Figure 1.

Comparison of LS mean (±SE) blood glucose levels in diabetes patients during high and low postprandial glucose conditions (A); between subject groups in high postprandial glucose conditions (B); and between subject groups in low postprandial glucose conditions (C). *, P < .001 high vs low postprandial glucose; †, P ≤ .01 diabetes patients with albuminuria vs nondiabetic controls; ‡, P < .001 diabetes patients with normal UAER vs nondiabetic controls. Whereas the T2D patients were studied during two periods (ie, under high and low postprandial glucose conditions), the nondiabetic controls were studied for only one period. The model-predicted mean data per time point for the control group were different between the two glycemic conditions due to the effect of adjusting for covariates.

Figure 2.

Comparisons of LS mean (±SE) brachial PWV, adjusted for age and BMI, between subject groups in high postprandial glucose conditions (A); between subject groups in low postprandial glucose conditions (B); and in diabetes patients during high vs low postprandial glucose conditions (C). *, Values ranged between P = .003 and P = .046, diabetes patients with albuminuria vs nondiabetic controls; †, P = .043, diabetes patients with normal UAER vs nondiabetic controls. Whereas the T2D patients were studied during two periods (ie, under high and low postprandial glucose conditions), the nondiabetic controls were studied for only one period. The model-predicted mean data per time point for the control group were different between the two glycemic conditions due to the effect of adjusting for covariates.

Figure 3.

Comparisons of LS mean (±SE) aortic PWV, adjusted for age and BMI, between subject groups, in high postprandial glucose conditions (A); between subject groups in low postprandial glucose conditions (B); and in diabetes patients during high vs low postprandial glucose conditions (C). Whereas the T2D patients were studied during two periods (ie, under high and low postprandial glucose conditions), the nondiabetic controls were studied for only one period. The model-predicted mean data per time point for the control group were different between the two glycemic conditions due to the effect of adjusting for covariates.