Association of Advanced Glycation End Products With Lower-Extremity Atherosclerotic Disease in Type 2 Diabetes Mellitus

Aims: Advanced glycation end products (AGEs) were reported to be correlated with the development of diabetes, as well as diabetic vascular complications. Therefore, this study aimed at investigating the association between AGEs and lower-extremity atherosclerotic disease (LEAD). Methods: A total of 1,013 type 2 diabetes patients were enrolled. LEAD was measured through color Doppler ultrasonography. The non-invasive skin autofluorescence method was performed for AGEs measurement. Considering that age plays an important role in both AGEs and LEAD, age-combined AGEs, i.e., AGEage index (define as AGEs × age/100) was used for related analysis. Results: The overall prevalence of LEAD was 48.9% (495/1,013). Patients with LEAD showed a significantly higher AGEage (p < 0.001), and the prevalence of LEAD increased with ascending AGEage levels (p for trend < 0.001). Logistic regression analysis revealed that AGEage was significantly positively associated with risk of LEAD, and the odds ratios of presence of LEAD across quartiles of AGEage were 1.00, 1.72 [95% confidence interval (CI) = 1.14–2.61], 2.72 (95% CI = 1.76–4.22), 4.29 (95% CI = 2.69–6.85) for multivariable-adjusted model (both p for trend < 0.001), respectively. The results were similar among patients of different sexes, body mass index, and with or without diabetes family history. Further, AGEage presented a better predictive value for LEAD than glycated hemoglobin A1c (HbA1c), with its sensitivity, specificity, and area under the curve of 75.5% (95% CI = 71.6–79.2%), 59.3% (95% CI = 54.9–63.6%), and 0.731 (0.703–0.758), respectively. Conclusion: AGEage, the non-invasive measured skin AGEs combined with age, seems to be a more promising approach than HbA1c in identifying patient at high risk of LEAD.


INTRODUCTION
Lower-extremity atherosclerotic disease (LEAD), defined as a buildup of fatty deposits in peripheral vascular (i.e., atherosclerosis) that leads to progressive narrowing of the lowerextremity arteries, is reported to be the primary manifestation of peripheral arterial disease (PAD) (1). Several researches have reported the close relationships between LEAD and cardiaccerebral vascular events (both non-fatal and fatal) including gangrene, amputation, and death (2,3). Diabetes is an established important risk factor for LEAD since the prognosis of LEAD is worse in patients with diabetes than those without (4,5).
LEAD may be silent or present with a variety of symptoms and signs indicative of extremity ischemia, such as claudication and rest pain (6,7). However, given the common concurrence of neuropathy in patients with diabetes, LEAD often remains clinically imperceptible until the symptoms become aggravated and advance to ulceration or gangrene due to the loss of pain sensation (8). Therefore, early identification and intervention of LEAD should be highlighted to delay its progression and effectively reduce the risk of the related adverse outcomes, thereby improving patients' quality of life.
Doppler ultrasound examination of atherosclerotic stenosis or occlusive lesions of the lower extremities is an importantly auxiliary method for LEAD diagnosis. However, considering the time-consuming and the risk of omissions of plaque due to widely distributed lower-extremity arteries, there is an urgent need for an indicator for early detection of early-stage LEAD, especially in patients with diabetes.
Advanced glycation end products (AGEs) are modifications of proteins or lipids that become non-enzymatically glycated and oxidized (9,10). AGEs affect nearly every type of cell and molecule in the body and are thought to be one factor in aging, as well as a causative role in diabetic vascular complications (11). Besides, considering the influence of age on LEAD, AGE age index, defined as AGEs × age/100, was constructed to investigate the association of AGE age with LEAD in patients with type 2 diabetes mellitus. In addition, whether AGE age can be used for early screening of patients at high risk of LEAD was also analyzed.

Study Population
Individuals diagnosed with type 2 diabetes who were admitted to the Department of Endocrinology and Metabolism, Shanghai Jiao Tong University Affiliated to Sixth People's Hospital during May 2017 to November 2019 were recruited. Type 2 diabetes mellitus were diagnosed based on 1999 World Health Organization (WHO) criteria (12). Inclusion criteria include (1) age ≥ 18 years with the presence of type 2 diabetes mellitus; (2) a stable glucoselowering regimen for the previous 3 months; (3) with valid data on both AGEs assessed by skin autofluorescence and lower limb ultrasound results; and (4) without any megascopic presence of dermopathy. Patients with diabetic ketoacidosis or severe and recurrent hypoglycemic events within the previous 3 months, and prior history of cardiovascular diseases, stroke, malignancy, mental disorders, or severe kidney, or liver dysfunction were excluded. Finally, 1,013 participants were enrolled into the final analysis.
This study was approved by the Shanghai Jiao Tong University Affiliated Sixth People's Hospital Ethics Committees and was in accordance with the Helsinki Declaration principles. Written informed consent was obtained from each participant.

Assessment of Covariates
Family history of diabetes, medical history, smoking status (current smoker or not), and current medication therapy including glucose-lowering drugs, antihypertensive drugs, lipidlowering drugs, and aspirin were recorded by self-report at baseline interview. Physical examination including height, weight, and blood pressure were performed in each patient. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Fasting venous blood sample was obtained after a 10-h overnight fasting. Biochemical measurements including glycated hemoglobin A 1c (HbA 1c ), glycated albumin (GA), fasting plasma glucose (FPG), fasting C-peptide (FCP), Creactive protein (CRP), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-c), and low-density lipoprotein cholesterol (LDL-c) were assayed as previously reported (13).

Assessment of AGEs
The spectroscopy device (Hefei Institutes of Physical Science, Chinese Academy of Sciences), mainly consisting of an ultraviolet light source, a broadband light source, a trifurcated fiber optic probe, and a compact charge coupled device spectrometer, was used to assess the skin AGEs. This device uses an excitation light with peak wavelength at 370 nm, which excites the AGEs in the skin that have fluorescence properties in a wavelength range of 420-600 nm. Besides, skin diffuse reflectance in a wavelength range of 350-600 nm was also detected to correct tissue absorption and scattering. The measurements were performed by trained nurses (at room temperature in a semidark environment) for three times at a normal skin site of the left volar side of the arm, and the average value was calculated for the analysis. AGE age was defined as AGEs × age/100.

Assessment of LEAD
Color Doppler ultrasonography was used for lower limb artery examination using an Acuson Sequoia 512 scanner (Siemens Medical Solutions, Mountain View, CA, USA) equipped with a linear array transducer with frequencies of 5-13 MHz. Seven arteries including femoral artery, deep femoral artery, superficial femoral artery, popliteal artery, anterior tibial artery, posterior tibial artery, and peroneal artery in each lower limb were measured for atherosclerotic plaque. The definition of atherosclerotic plaque have been described in detail previously, i.e., a focal structure encroaching into the arterial lumen ≥ 0.5 mm, 50% of the surrounding intima-media thickness (IMT) value, or an IMT thickness ≥ 1.5 mm (14)(15)(16). The presence of atherosclerotic plaques in any of the lower limb artery segments listed above was defined as LEAD (17). model was used to examine the independent influencing factors of AGE age . Binary logistic regression analysis was used to assess the association between quartiles of AGE age and LEAD. The receiver operating characteristic curve was used to evaluate the efficacy of AGE age and HbA 1c in early detection of LEAD. A restricted cubic spline nested in logistic models was performed to test whether there was a dose-response or non-linear association of AGE age as a continuous variable with the odds of LEAD. A two-tailed p-value of < 0.05 was considered statistically significant.
Besides, we also found an increase in AGE age in patients with LEAD compared with those without LEAD [49.3 (43.2-56.5) vs. 41.2 (32.5-48.8), p < 0.001]. Similarly, as shown in Figure 1, the prevalence of LEAD increased progressively across the categories of increasing AGE age (p for trend < 0.001). Then, the participants were stratified according to AGE age levels (AGE age < 43.2 and AGE age ≥ 43.2). Supplementary Table 1 depicts the characteristics of subjects by AGE age levels.
Multivariate linear regression analysis defined the AGE age levels as the dependent variable, and sex, diabetes duration, BMI, SBP, family history of diabetes, HbA 1c , FPG, lipid profile, CRP, smoking status, antidiabetic therapy, antihypertensive medication, lipid-lowering medication, and aspirin use were designated as the independent variables. The results showed an independently positive association between diabetes duration, SBP, HDL-c, antihypertensive medication, lipid-lowering medication, and AGE age (all p < 0.01). Besides, BMI and TC were negatively associated with the AGE age levels (all p < 0.001).
When AGE age was considered as a continuous variable by using restricted cubic splines, a graded positive association of AGE age with the odds of presence of LEAD was observed (p for trend < 0.001; Figure 2). This curve trend was consistent with the findings in Table 2 when AGE age was considered as a categorical variable.
When stratified by sexes, BMI, HbA 1c , never and past or current smokers, as well as taking antidiabetic, lipidlowering, antihypertensive, and antiplatelet medication or not, the graded positive association between AGE age and the odds of presence of LEAD were consistent across all subgroups. Subgroup analyses were then performed to examine potential effect modifiers, and we observed an interaction of HbA 1c and using antiplatelet medication or not with a p for interaction <0.01 (Supplementary Table 2). Supplementary Table 3 described the confounding factors associated with LEAD.
Areas under the curve (AUCs) of AGE age and HbA 1c for early detection of LEAD were measured. The results showed that AGE age had a better predictive value for LEAD than HbA 1c , concretely, 0.731 (0.703-0.758) for AGE age and 0.513 (0.482-0.544) for HbA 1c , respectively (p < 0.01). The optimal cutoff point for AGE age in early detecting LEAD was 43.2, with a sensitivity of 75.5% (95% CI = 71.6-79.2%) and a specificity of 59.3% (95% CI = 54.9-63.6%). The study participants were stratified based on sex, BMI, diabetes family history, and current smoker or not, and the results showed an acceptable efficacy of AGE age < 43.2 in early identifying LEAD in all related subgroups. In addition, the efficacy of AGE age in early detection of LEAD seems more pronounced in male and BMI < 25 kg/m 2 subgroup type 2 diabetic patients ( Table 3).

DISCUSSION
In this observational study, the AGE age index was proposed for the first time. We observed a positive association of AGE age with the odds of presence of LEAD among patients with type 2 diabetes mellitus, independent of traditional risk factors for LEAD that include HbA 1c . Besides, AGE age may be a suitable indicator for early identification of patients at high risk of LEAD, with its optimal cutoff point of 43.2. LEAD is an emerging public health burden with an endemic progression worldwide that affects over 200 million people worldwide (18). Related studies reported that LEAD was two to four times more frequent in people with type 2 diabetes than in the general population (19,20), and the prevalence of LEAD also increased along with the rising diabetes duration as shown in the United Kingdom Prospective Diabetes Study (UKPDS), concretely, 1.2% when first diagnosed with diabetes, and increased up to 12.5% after 18 years of evolution (21). Besides, LEAD is particularly frequent in diabetic patients with worse outcomes, especially the risk of lower limb amputation, four to five times higher, compared with non-diabetic subjects (22,23), suggesting that poor glycemic control may play an important role in LEAD progression.
HbA 1c is a well-established marker for assessment of glycemic control. In the UKPDS trial, each 1% reduction in HbA 1c was associated with a 43% reduction in the risk of major LEAD (amputation or death induced by peripheral vascular event) (24). However, HbA 1c is insufficient in terms of the overall evaluation of glycemic control, i.e., patients with similar HbA 1c could have totally distinct glucose profiles. The results of the ADVANCE trial demonstrated that the incidence of major LEAD (lowerlimb ulceration, amputation, revascularization requirement, or death following a PAD) was comparable between intensive and standard glucose control groups (25,26). Moreover, type 2 diabetic patients with or without LEAD in the current study share similar HbA 1c levels. Therefore, other factors related to glycemic control beyond HbA 1c may be related to LEAD.
AGEs are considered as one factor in aging and some age-related chronic diseases including Alzheimer's disease, cardiovascular disease, stroke, and diabetes. Studies have reported that hyperglycemic status may promote the accumulation of AGEs, while AGEs can cause vascular stiffening and entrapment of LDL particles in the artery walls by inducing crosslinking of collagen in the context of cardiovascular disease (27,28). AGEs can also cause LDL glycation, thereby further promoting its oxidation, while oxidized LDL is one of the major factors in the development of atherosclerosis. In addition, the interaction between AGEs and RAGE (receptor for AGEs) induced oxidative stress, and activation of inflammatory pathways in vascular endothelial cells also plays an important role in the development of systemic atherosclerosis including LEAD (28). These findings raised the possibility that AGEs and AGE-generated indicators may be an alternative indicator reflecting LEAD.
Considering the fact that AGEs are closely correlated with age, while LEAD is reported to be discovered during the fifth decade of life and the prevalence of LEAD increased exponentially after 65 years (29), we proposed the AGE age index for the first time, which combines AGEs and age organically. Consistent with our hypothesis, in the current study, the level of AGE age was significantly increased in LEAD patients with type 2 diabetes mellitus. Accordingly, we also found a graded positive association of AGE age with the odds of presence of LEAD, even after adjusting for clinical risk factors, including HbA 1c , which imply the value of AGE age in assessing the risk of diabetic complications independent of HbA 1c . Particularly, skin autofluorescence is a non-invasive method for AGEs detection. All the above-mentioned suggest that AGE age may be a simple and effective indicator for predicting LEAD.
Meanwhile, we proposed that in type 2 diabetes patients, AGE age might be a more suitable indicator than HbA1c for mimicking the poor prognosis of diabetes, i.e., LEAD, regardless of gender, BMI, and family history of diabetes. Our results showed that AGE age had a significantly higher predictive value for LEAD than HbA 1c . With the cutoff point of 43.2, around 3/4 patients with LEAD (384/591) were successfully identified. Based on the current study, we recommend type 2 diabetic patients whose AGE age ≥ 43.2 are considered at high risk for LEAD that needs ultrasound for further confirmation.
The relatively large sample size and well-documented clinical information of the current study makes our findings more reliable. Nevertheless, there are some limitations that should be pointed out. First, since the current study was a cross-sectional study, the cause-and-effect relationship between AGEs and LEAD could not be clarified. Second, the participants enrolled in the current study were Chinese type 2 diabetic hospitalized patients. Considering the racial difference in circulating AGEs, as well as the promotion of hyperglycemia on AGE accumulation, whether our results can be generalized to all diabetic patients and even patients with cardiovascular diseases needs further study.

CONCLUSION
In conclusion, we provide evidence that AGE age is associated with the prevalence of LEAD in patients with type 2 diabetes mellitus independent of HbA 1c . AGE age , the non-invasive measurement of accumulated AGEs combined with age, seems a promising approach than HbA 1c to mimic the poor prognosis of hyperglycemia, i.e., triage for patient at high risk of LEAD.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Shanghai Jiao Tong University Affiliated Sixth People's Hospital Ethics Committees. The patients/participants provided their written informed consent to participate in this study.