Diabetes as well as its complications has become increasingly serious public health burden. Peripheral vascular disease, heart disease, and stroke are all highly prevalent in diabetes patients, and traditional microvascular complications such as retinopathy, nephropathy, and neuropathy also occur frequently . ABI was proposed by Winsor et al. in 1950 and was initially used for the non-invasive diagnosis of lower limb PAD, which is a kind of diffuse atherosclerotic vascular disease and has a prevalence ranging from 21.1% to 38.5% in > 30 year-old diabetic patients [18,19,20], increasing with age and the presence of cardiovascular risk factors . Studies have shown that diabetes is the main risk factor for the onset of PAD . Later, ABI has become an indicator of vascular atherosclerosis, as well as a prognostic indicator of cardiovascular events and dysfunction . Low ABI is a major risk factor for cardiovascular morbidity and mortality in diabetic patients [22, 23], and higher-than-normal ABI also significantly predicts subsequent all-cause mortality . Multiple studies have confirmed that abnormal ABI is associated with an increased risk of cardiovascular or all-cause mortality and major adverse cardiovascular events in diabetes patients. Subgroup analysis showed that abnormally low and high ABI had similar values in predicting cardiovascular mortality  and there was a U-shaped association between ABI values and cardiovascular mortality . Although PAD is common in diabetics, it’s still not fully recognized, and its diagnosis is often difficult when patients have peripheral neuropathy, as the latter condition can mask pain . Therefore, defining the etiology, investigating risk factors, and exploring possible early detection and treatment strategies are essential for PAD prevention and control.
Numerous studies have shown significant clinical benefits of using CGM in patients with diabetes, regardless of insulin administration . CGM provides a full range of glucose parameters including TIR, which fills the blind spot of HbA1C and traditional SMBG monitoring and is one of the future trends of diabetes metabolic monitoring. As a new indicator, TIR is a simple and intuitive metric that provides information about the quality of glucose control, reporting the time at which blood sugar levels reached a given target during a given period, allowing personalized analysis, and taking into account individual factors such as diabetes type, age, pregnancy status, and complications. A recent study of 4268 patients found that optimal TIR was the greatest motivator for patients with type 1 diabetes to voluntarily choose a treatment . Another online survey of 3455 patients with diabetes found that TIR was the primary indicator of most concern in patients with type 1 diabetes and type 2 diabetes with or without insulin therapy. In addition, TIR better reflected the effect of acute glucose intervention than HbA1c alone , suggesting its possible higher value than HbA1C in evaluating treatment options. The 2017 International Consensus on Use of Continuous Glucose Monitoring recommended TIR as one of the 14 key indicators to be included in the CGM standard report . In 2019, the ATTD panel updated its recommendations in the International Consensus on TIR, positioning TIR as one of the most important indicators in diabetes management and reaching a consensus on glycemic cutoff points and target timing for personalization in different groups of patients . In summary, TIR has become a key indicator for assessing glucose control, a simple and intuitive reflection of the acute effects of treatment and lifestyle changes, and a convenient tool to help patients better understand and manage their condition.
In clinical trials, TIR has been shown to be negatively associated with HbA1c and associated with the risk of long-term diabetic complications as an endpoint. Regarding diabetic microvascular complications, relevant studies have confirmed that TIR has significantly negative correlations with the risk of all stages of diabetic retinopathy, microalbuminuria outcomes [31, 32], proteinuria , and diabetic peripheral neuropathy [34,35,36]. As for macrovascular complications of diabetes, TIR is significantly negatively associated with the risk of abnormal carotid artery intima-media thickness  and long-term all-cause and cardiovascular mortality . In our study, with the increase of TIR, the prevalence rate of diabetic retinopathy, ACR, HbA1C(%), SD, MODD, MAGE, MBG, ADDR, M value, and CV showed a decreasing trend (P < 0.05) in type 2 diabetes patients, which was consistent with previous studies.
The purpose of this study was to explore the relationship between TIR and ABI anomalies. The level of TIR in the abnormal ABI group was lower than that in the normal group (P < 0.05). With the increase of TIR quartiles, the prevalence rate of ABI abnormalities decreased significantly at P < 0.05. Spearman correlation analysis showed that TIR was negatively correlated with ABI (P = 0.002). In logistic regression, TIR (OR: 0.979, 95% CI: 0.967–0.990, P < 0.001) was significantly associated with the presence of abnormal ABI in the unadjusted model. The correlation persisted after adjustment for cardiovascular risk factors and other indicators including age, diabetes duration, sex, blood pressure, lipid profiles, UREA, Scr, ACR, BMI, HbA1C(%), and GV indicators (P < 0.05). Based on our results, we found a significant association between TIR and ABI abnormalities independent of HbA1C. More importantly, lower TIR was associated with the presence of abnormal ABI even after adjusting for GV indicators and cardiovascular risk factors, suggesting that the value of TIR in predicting the risk of ABI abnormalities was independent of GV indicators and other cardiovascular risk factors. The significant association of TIR with ABI revealed a potential association between TIR and PAD and arterial calcification, further supporting TIR as a valuable glucose indicator and a reasonable clinical outcome in scientific research.
As a classical glucose monitoring indicator, HbA1C has been widely used in clinical and scientific research worldwide since the 1990s. Its correlation with TIR has been well studied, and it was estimated that there is a 0.8% (9 mmol/mol) change in HbA1C for every 10% change in TIR [15, 32], which was also confirmed in our study. The association between HbA1C and ABI has also been explored. Some cross-sectional observational studies showed that HbA1C and ABI are correlated, with high HbA1C being an independent risk factor of low ABI (ABI ≤ 0.9) ; and patients with HbA1C ≥ 7% had 2.9 times the risk of microalbuminuria ( +) and ABI ≤ 0.90 compared with patients with HbA1C < 7% (P = 0.043 and 0.048, respectively) . However, some studies have come to different conclusions. Studies conducted by Papazafiropoulou et al.  and Sayilan et al.  showed no correlation between ABI value and HbA1C in type 2 diabetes patients, and Zhengliang et al. showed no correlation between ABI and HbA1C in the Shanghai elderly population (including the diabetic group and control group) . In our study, HbA1C also failed to indicate ABI abnormalities. There was no significant difference in HbA1C between different ABI groups, and the correlation between HbA1C and abnormal ABI was not statistically significant (P = 0.623). Binary logistic regression showed that there was no correlation between HbA1C and abnormal ABI in all models. We suggest that TIR may be a more valuable clinical indicator than HbA1C in indicating ABI abnormalities, but further studies are needed to confirm this conclusion. One possible explanation is that HbA1C is primarily associated with microvascular diseases, while ABI is primarily associated with macrovascular complications such as myocardial infarction, stroke, or PAD . On the other hand, HbA1C cannot accurately reflect hypoglycemia and blood sugar fluctuations, which may also be risk factors for diabetic vascular diseases [45,46,47]. Therefore, TIR has an advantage over HbA1C in assessing the risk of ABI anomalies.
In fact, for individuals, elevated HbA1C levels do not provide clinicians with specific recommendations for adjusting treatment regimens, and studies have shown considerable inter-individual differences in average blood glucose levels even when patients have normal HbA1C levels. HbA1C is unable to well distinguish the HbA1C components generated by physiological and pathological blood glucose exposure and fluctuation, and cannot reflect the degree of blood glucose fluctuation. Compared with HbA1C, TIR has certain advantages. Firstly, TIR is influenced by all factors affecting daily glucose patterns, and it collects all glucose level data within a given time range and can reflect glucose fluctuations. When HbA1C levels do not reflect hypoglycemia, it may result in false good HbA1C levels in such circumstances. Secondly, TIR is more accessible and intuitive, enabling patients to know about their blood glucose control level more clearly, encouraging patients to control their diabetes, and helping them improve their blood glucose control with real-time data. Thirdly, TIR is a more accurate assessment of glycemic control than HbA1C when HbA1C is inconsistent with mean glucose, in conditions such as iron deficiency or other anemia, hemoglobulin disease, and pregnancy . Some clinical researchers believe that with the accumulation of more evidences and the advancement of CGM research and development, TIR is expected to surpass HbA1C in the future and become one of the main evaluation indicators for blood glucose control and management .
According to our results, with the decrease of TIR, SD, MODD, MAGE, MBG, ADDR, M value, and CV of type 2 diabetes patients increased (P < 0.05), that is, the glycemic variability as well as hypoglycemia and hyperglycemia events increased. Some studies have suggested that vascular endothelial dysfunction is considered to be the key pathogenic basis of type 2 diabetes macrovascular complications, and hyperglycemia may cause vascular endothelial dysfunction and accelerate the occurrence of diabetic macrovascular complications . On the side, glycemic variability exacerbates oxidative stress in type 2 diabetes patients, further damages endothelial cells, and leads to the occurrence of diabetic macrovascular complications through increased inflammation or epigenetic changes [46, 47]. These may partly explain why TIR is associated with the macrovascular complications suggested by ABI abnormalities, but the specific mechanisms need to be further investigated.
Previous studies have shown a J-shaped association between the ABI and diabetic mortality and the occurrence of vascular complications, with the risk increasing in the low ABI population and continuing to increase as the ABI decreases , but studies on high ABI are still scarce. As for the abnormal changes of ABI values of both low and high in the same state of low TIR, we speculate that there may be the following reasons. Firstly, glucose control level represented by TIR is only one of the main influencing factors for PAD and arterial calcification, and the two diseases are also influenced by various other factors, such as advanced age, diabetes duration, smoking, hypertension, hyperlipidemia, etc. And even if they share the same influencing factors, the contribution to the disease of these factors may be different. In our study, two groups were compared on the premise that there were significant differences among the three ABI groups. On the premise that there were no significant differences among other groups in the same index, only LDL values in the low and high ABI groups showed significant difference (P = 0.009), and LDL in the low ABI group was significantly higher than that in the high ABI group, suggesting that the degree of LDL-mediated atherosclerosis is the most important contributing factor of low ABI, and preliminary indicating that type 2 diabetes patients with higher LDL are more prone to have lower ABI. There was a significant difference in Scr between normal and low ABI groups (P = 0.007), and there were significant differences in TIR (P = 0.003), age (P = 0.023), UREA (P = 0.006), ACR (P = 0.004), TAR (P = 0.015), and MBG (P = 0.014) between normal and high ABI groups, and in diabetes duration between both normal and low (P = 0.023) and normal and high (P = 0.006) ABI groups. We can infer that elevated Scr may be a risk factor for low ABI (consistent with previous studies ); advanced age, high UREA, high ACR, and hyperglycemia are more likely to be risk factors for high ABI; and long diabetes duration may be a risk factor for both.
Low ABI has been shown to be associated with many cardiovascular risk factors. Previous studies have shown that the most serious risk factors for PAD were diabetes and smoking; others included advanced age, hypertension, and hyperlipidemia; potential risk factors included CRP, fibrinogen, homocysteine, apolipoprotein B, lipoprotein (a), and elevated plasma viscosity. In multivariate analyses of most studies, TC and low HDL levels were associated with PAD. Genetics, poverty, environmental pollution, and physical inactivity have also been linked to PAD . In diabetics, the risk of PAD increased with age, diabetes duration, insulin use, and the presence of peripheral neuropathy ; every 1% increase in HbA1C was associated with a 26–28% increase in the risk of developing PAD, according to the UK Prospective Diabetes Study (UKPDS) [51, 52]. Eraso et al. carried out a study on PAD and prevalence and cumulative risk factor spectrum analysis, and Joosten et al. carried out a study on the relationship between conventional cardiovascular risk factors and male PAD risk factors, and both found that the impact of risk factors is cumulative, and the more the number of risk factors, the greater the risk of PAD [24, 53].
On the other side, studies have shown that high ABI is directly associated with male sex, diabetes, hypertension, BMI, and age, but negatively associated with smoking and hyperlipidemia, with diabetes being the strongest risk factor [6, 54]. Patients with clinical neuropathy, nephropathy, and long diabetes duration were considered to be at high risk for arterial calcification . Genome-Wide Association Study has identified multiple contributing sites associated with atherosclerosis, diabetes, and coronary artery calcification; high glucose levels were observed to be highly associated with vascular calcification and vascular disease . Risk factors for vascular and valve calcification included aging, metabolic syndrome, smoking, and male sex . In addition, according to the pathophysiological process of vascular calcification, its driving factors included elevated serum phosphate, advanced glycation end products, bone morphogenetic protein, inflammatory cytokines, and leptin. Magnesium, antioxidants, vitamin K, and sufficient but not excessive vitamin D status seemed to prevent arterial calcification .
In summary, different risk factors for low and high ABI suggest that the pathogenesis of PAD and arterial calcification may differ. And differences in study populations may explain differences between study results.
On the other hand, arterial calcification may cover PAD and lead to a relatively abnormally high ABI. Although the two diseases often coexist, when vascular calcification is present, the ABI cannot detect stenosis due to the decreased compressibility of the arteries . High ABI values are associated with atherosclerosis secondary to arterial calcification and may lead to an underestimation of the prevalence of PAD in diabetes in cases where complex or long-term diabetes leads to more arterial calcification. In fact, among diabetics with high cardiovascular risk and neuropathy and with normal ABI between 0.9 and 1.3, the prevalence of PAD as measured by DUS was 57% ; other authors have also reported a high prevalence of PAD in diabetic patients with elevated ABI, estimated to be between 58 and 84% [59, 60]. Researchers suggested that reduced blood flow to the lower extremities in diabetics due to arteriosclerosis may explain the association between high ABI and PAD . In such cases, it is recommended that a duplex ultrasound must be performed to confirm and assess PAD .
There are some limitations of our study. Firstly, it was a single-center study with a relatively small sample size, which may lead to limited statistical certainty. Secondly, studies showed that the CGM of 70% of patients in the last 14 days was closely related to 3-month mean glucose, TIR, and hyperglycemia. Our study conducted CGM for only 72 h, which may not be so sufficient . Thirdly, as a cross-sectional study, our study only described the correlation between TIR and ABI abnormalities, but could not provide more information about the causal relationship. Fourthly, single rather than dynamic measurement of the ABI may lead to individual selection bias. Finally, arterial occlusion or calcification had not been confirmed by diagnostic imaging, such as angiography or ultrasound, and computed tomography. In these patients, ABI values may tend to be pseudo-normalized, leading to the misclassification of ABI categories.