Cardiovascular disease (CVD) has the highest mortality rate worldwide [1], and vascular stenotic and occlusive lesions are the main pathological cause of death and disability for CVD patients [2]. At present, the development of intravascular therapy technology provides a reliable method for the treatment of vascular occlusive diseases, and quite a few of the difficulties in treatment have been resolved. However, restenosis after endovascular repair is a major problem that confuses clinicians and affects their choice of treatment [3]. Studies have shown that 30% to 50% of patients with coronary ischemic disease experience restenosis after endovascular therapy within five years; although the application of drug-coated balloons or drug-eluting stents has clearly reduced the occurrence of restenosis, restenosis still occurs in 10% to 20% of patients within one year [4, 5]. Restenosis after endovascular treatment causes a large number of patients to stop working or even die, which causes considerable damage to people's health and social and economic development. Thus, this is an urgent unsolved clinical problem. Identifying potential exposure and protective factors could make health care more effective in controlling restenosis after endovascular treatment.

Diabetes mellitus (DM) and its complications constitute one of the greatest human health problems worldwide. The prevalence of DM will increase globally from 371 million individuals in 2013 to 552 million individuals in 2030 [6]. DM is an important risk factor for the development of atherosclerotic diseases such as coronary heart disease (CHD), cerebrovascular disease, and peripheral artery disease (PAD). Cardiovascular complications are the leading cause of mortality among individuals with DM, and > 50% of patients die from a cardiovascular event, especially coronary artery disease but also stroke and peripheral vascular disease [7]. Insulin resistance and hyperglycemia in diabetes patients increase the risk of adverse cardiovascular events [8]. Studies have shown that hyperglycemia, insulin resistance, and an increase in advanced glycation end products are important conditions for a 2–fourfold increased risk of coronary artery disease (CAD) and PAD among people with diabetes [9,10,11,12,13,14]. However, at present, there is insufficient clinical evidence to confirm that diabetes increases the risk of restenosis after endovascular treatment. Therefore, to determine the effect of DM on restenosis among patients following endovascular interventional therapy, we designed this meta-analysis based on current clinical randomized controlled trial (RCT) data.

Methods

Study design and search strategy

We have designed the research and register for the study on the INPLASY website, and the registration number is INPLASY202370034 (DOI: 10.37766 / inplasy2023.7.0034). Ethical approval was acquired from the Ethic Committee of Southwest Medical University (No. 20220217–013). Two researchers independently scanned the titles and abstracts of the retrieved studies for the topic, and then obtained the full texts of potentially eligible studies and examined these independently for their suitability according to the inclusion criteria. In the case of disagreement between the two researchers, a third researcher was consulted to reach a consensus on whether to include the report or not. They documented the selection process with a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA [15]) flow chart. Trials were included if they satisfied the following eligibility criteria: (1) The studies included had to be randomized controlled studies (RCTs) including patients with or without DM; (2) had to involve lesions confined to the coronary arteries or femoral popliteal artery; (3) had to involve endovascular treatment via percutaneous transluminal angioplasty (PTA) or stenting; and (4) had to include an outcome of restenosis at the target lesion site. The exclusion criteria were as follows: (1) the proportion of patients lost to follow-up was higher than 20%; (2) a secondary restenosis operation; and (3) a sub-analysis or post-hoc analysis of RCTs. We developed and adhered to a standard protocol for study identification, inclusion, and data abstraction for all steps of our systematic review. Our major endpoint was restenosis, defined as a stenosis diameter > 50% in the in-segment area assessed by angiography, including the stent area as well as 5-mm margins proximal and distal to the stent. Meanwhile, clinically driven target lesion revascularization (CD-TLR) was also included, defined as any procedure performed to restore luminal patency after there has been late luminal loss of the target lesion (confirmed by angiography).

In this meta-analysis, we identified related published studies through a computerized literature search of the PubMed/Medline, Cochrane Library, and EMBASE (from 01/1990 to 12/2022) electronic databases. Two independent researchers checked citations for inclusion in the systematic review by using a hierarchical approach. The researchers assessed the title, abstract, and full text of these manuscripts. In addition, another reviewer manually searched the bibliographies of journal articles and relevant reviews to locate additional studies.

Reviewers extracted data using a data extraction form designed and piloted by the authors. If studies were reported in multiple publications, data were extracted from the different publications and then combined into a single data extraction form so that no data were omitted. The following characteristics of the included studies were registered in the data extraction form: methods and study design, participants, interventions and outcomes, including the outcome of restenosis. For all unclear restenosis analysis results, emails were sent to the corresponding authors to request the raw data; unfortunately, up to the time that the manuscript was written, no reply had been received.

Data synthesis and statistical analysis

For dichotomous data, we calculated Mantel‒Haenszel relative ratios (RRs) and 95% confidence intervals (CIs). Heterogeneity across studies was assessed by Cochran’s Q statistic with a P value set at 0.1. The I2 statistic was also taken into account regardless of the P value. An I2 of ≥ 50% was prespecified as the threshold considered too high to provide consistent analysis. A random effects model was used for the analysis. Tests were two-tailed, and a P value of < 0.05 was considered statistically significant. Funnel plots were used to assess publication bias. STATA 12.0 (StataCorp, USA) was used to analyze the data.

Assessment of the risk of bias in the included studies

Two reviewers assessed the risk of bias independently for the included studies using the Cochrane risk of bias assessment tool [16]. We evaluated all included studies for the following: adequacy of sequence generation and allocation concealment; adequacy of blinding of couples, providers and outcome assessors; completeness of outcome data; risk of selective outcome reporting; and risk of other potential sources of bias. The results of the risk of bias assessment are presented in Fig. 2.

Results

The results of the literature search

In this analysis, a total of 2,996 RCTs were retrieved from the database. After eliminating duplicate studies, 1,430 studies remained. After browsing titles and abstracts, 605 full-text articles were assessed for eligibility, of which 585 were excluded because of the absence of relevant endpoint data (no surveillance undertaken, restenosis rates not reported separately). Finally, 20 studies were included in the meta-analysis for qualitative and quantitative analyses (Fig. 1). As a result, 31,066 patients were enrolled in this study. Table 1 details the case numbers and baseline characteristics, inclusion/exclusion criteria, number and type of stent/PTA procedure, strategies for follow-up, criteria for diagnosing restenosis, and the number of cases of restenosis with DM for each RCT. All authors of the studies included in this meta-analysis were requested to supply missing data and details of their studies. Unfortunately, no author supplied the requested information.

Fig. 1
figure 1

Flow diagram of the study selection process

Full size image
Table 1 Baseline characteristics and quality assessment of the included studies
Full size table

Description of the prospective randomized pooled trials and quality assessment of the included studies

As the aim of this study was to analyze the PTA- or stenting-related outcomes among patients who suffered from cardiovascular stenotic or occlusive disease with or without DM, the efficacy endpoint was the rate of restenosis (> 50% stenosis at angiography or TLR during the follow-up period). As a result, 31,066 patients were enrolled in this study. Table 1 summarizes the characteristics of the 20 included RCTs. All trials met the inclusion criteria and had a low risk of bias according to the Cochrane tool [16] for assessing risk of bias in RCTs (Fig. 2). Four studies (TOSCA, SORT OUT III, ICE, HORIZONS AMI) were designed to be open-labeled, and the BASKET-SMALL 2 study did not describe its randomization method (Table 2).

Fig. 2
figure 2

Risk of bias evaluated by the RoB 2 tool. All trials met the criteria and had a low risk of bias according to the Cochrane tool for assessing the risk of bias in RCTs

Full size image
Table 2 Blood glucose control and evaluation form
Full size table

Meta-analysis results

There were 20 RCTs that included a total of 31,066 patients that reported data on DM and restenosis. Since there was high heterogeneity (I2 = 57.3%, P = 0.001), a meta-analysis was conducted through a random effects model. Overall, the pooled results showed that DM was significantly associated with a higher risk of a major endpoint (RR = 1.43, 95% CI: 1.25–1.62; Fig. 3). The heterogeneity test showed significant differences among individual studies (P < 0.01, I2 = 57.3%). The sensitivity analysis showed that heterogeneity mainly came from BIONICS (2018), but the outcome did not change with removal of this study (RR 1.40,95% CI: 1.23–1.59) (Fig. 9).

Fig. 3
figure 3

Forest plot of the association between DM and restenosis. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 3 (95% CI, 1.25–1.62; I2= 57.53%, P = 0.001). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image

Subgroup analysis

To determine the relationship between glycemic control levels and the incidence of postoperative restenosis, we performed a subgroup analysis of the included studies. Since detailed blood glucose levels of patients could not be obtained, we divided the included studies into the "under control" group and the "unknown" group according to whether the proportion of overall medicine glycemic control (oral hypoglycemic agents or insulin) of diabetic patients in each study was more than 80% (Fig. 4). As a result, there was no significant difference in restenosis rates between the under-control group (RR = 1.53, 95% CI 1.26–1.84, P < 0.05) and the unknown group (RR = 1.37, 95% CI 1.16–1.63, P = 0.001).

Fig. 4
figure 4

Subgroup analysis of the association between restenosis and glycemic control levels. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 4 (‘Under Control’ subgroup, RR = 1.53, 95% CI, 1.25–1.62; I2 = 36.9%, P < 0.05, ‘Unknown’ subgroup, RR = 1.37, 95% CI, 1.16–1.63; I2 = 63%, P = 0.001.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image

Antiplatelet therapy and anticoagulant therapy after vascular intervention are closely related to long-term efficacy [17]. Studies have demonstrated that an adequate treatment course of antiplatelet therapy after coronary stenting or angioplasty of lower extremity peripheral arteries is beneficial to improve long-term patency rates [18, 19]. We also analyzed the relationship between the postoperative dual antiplatelet therapy (DAPT) duration and restenosis in the studies. The results showed that regardless of the duration of DAPT, either less than (RR = 1.53, 95% CI 1.23–1.90, P < 0.05) or more than 6 months (RR = 1.48, 95% CI 1.27–1.73, P = 0.006) had no effect on restenosis outcome (Fig. 5). In addition, subgroup analysis was performed based on a 1-year follow-up. The results showed that there was no significant difference in the incidence of restenosis based on follow-up duration in the more than 1 year group (RR = 1.41, 95% CI 1.16–1.70, P = 0.018) and the less than or equal to 1 year group (RR = 1.44, 95% CI 1.19–1.74, P = 0.003) (Fig. 6). Furthermore, in these 20 studies, the major target vessel was the coronary artery, except for the ICE study, which evaluated peripheral arteries. Then, we performed a subgroup analysis by different target lesions. As shown in the forest plot, the restenosis rate after endovascular therapy was related to the primary site of the vascular lesion (coronary artery (RR = 1.44, 95% CI 1.27–1.64) or lower limb artery (RR = 0.75, 95% CI 0.33–1.74)) (Fig. 7). Moreover, subgroup analysis was also performed to account for continental differences in the included populations. Since there was only one study conducted among Asians, our subgroup included European and American participants (without distinguishing between South and North America), and it showed no significant differences in the endpoints between Americans (RR = 1.53, 95% CI 1.22–1.92, P = 0.004) and Europeans (RR = 1.43, 95% CI 1.21–1.69, P < 0.05) (Fig. 8). According to the diagnosis of restenosis, including angiography and TLR, we conducted a subgroup analysis of the two modalities, and the results showed that patients with diabetes had a higher risk of TLR (RR = 1.53, 95% CI 1.34–1.76, P = 0.010) (Fig. 9).

Fig. 5
figure 5

Subgroup analysis of the association between restenosis and the duration of DAPT. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 5 (‘More than 6 months’ subgroup, RR = 1.48, 95% CI, 1.27–1.73; I2 = 56.4%, P = 0.006, ‘Equal to or less than 6 months’ subgroup, RR = 1.53, 95% CI, 1.23–1.90; I2 = 18.2%, P < 0.05.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image
Fig. 6
figure 6

Subgroup analysis of the association between restenosis and different follow-up times. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 6 (‘More than 1 year’ subgroup, RR = 1.41, 95% CI, 1.16–1.70; I2 = 58.7%, P = 0.018, ‘Equal to or less than 1 year’ subgroup, RR = 1.44, 95% CI, 1.19–1.74; I2 = 60.8%, P = 0.003.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image
Fig. 7
figure 7

Subgroup analysis of the association between restenosis and types of lesion vessels. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 7 (‘Coronary’ subgroup, RR = 1.44, 95% CI, 1.27–1.64; I2 = 57.8%, P = 0.001, ‘Peripheral’ subgroup, RR = 0.75, 95% CI, 0.33–1.74.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image
Fig. 8
figure 8

Subgroup analysis of the association between restenosis and different continents. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 8 (‘Europe’ subgroup, RR = 1.43, 95% CI, 1.21–1.69; I2 = 49.6%, P < 0.05. ‘America’ subgroup, RR = 1.53, 95% CI, 1.22–1.92; I2 = 68.2%, P = 0.004.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image
Fig. 9
figure 9

Subgroup analysis of the association between restenosis and diagnostic methods. The vertical dashed lines indicate the pooled summary estimate (95% CI) for all studies in Fig. 9 (‘TLR’ subgroup, RR = 1.56, 95% CI, 1.35–1.80; I2 = 49.6%, P = 0.008, ‘Angiography’ subgroup, RR = 1.12, 95% CI, 0.94–1.33; I2 = 15.4%, P = 0.315.). The area of each square is proportional to the inverse variance of the estimate. The horizontal lines indicate the 95% confidence intervals of the estimate

Full size image

Quantified covariable analysis-meta-regression analyses

With the aim of performing a comprehensive literature review on restenosis after interventional endovascular treatments of PTA or stenting in patients with diabetes (1) across different “interventions” (we performed a regression analysis according to different interventional modalities, i.e., aspirin, PTA or stent placement) (2) among patients with different health conditions (the proportion of patients with smoking exposure, hypertension, or hyperlipemia, was distinguished and regression analysis was performed) (3) and in different periods of these interventional endovascular techniques (presented as the publication times), meta-regression was employed. RRs, using variable rates as the dependent variable, and the different interventions, the different health conditions, and the publication times as the independent variables, were determined. There was no evidence that the different interventions, the different health conditions and the publication times were confounding factors in this subgroup analysis. Data from the analyses of moderator variables are presented in Table 3.

Table 3 Univariate meta-regression for restenosis
Full size table

Sensitivity analysis

In the sensitivity analysis, each included study was removed one by one, and a summary analysis of the remaining studies was performed to assess whether a single included study had an excessive impact on the results of the entire meta-analysis (Fig. 10). None of the studies had an excessive impact on the results of the meta-analysis, indicating that the results of the meta-analysis were stable and reliable.

Fig. 10
figure 10

Sensitivity analysis of the association between diabetes mellitus and the endpoints

Full size image

Publication bias

The probability of publication bias in the spread of the meta-analysis by funnel diagram and Begg’s test at a significance level of 0.05 indicated no bias of spread in the present study (p = 0.344) (Fig. 11). According to the results of the diagram, the publication offset of the included studies was small, and the results of the meta-analysis had high uniformity.

Fig. 11
figure 11

Funnel plot for assessing publication bias

Full size image

Meta-analysis GRADE assessment

The evidence was assessed according to the GRADE process for the purposes of making clinical practice recommendations. We used GRADE to evaluate the quality of evidence, as shown in Fig. 12. Judgments about evidence quality (high, moderate, low or very low) were made by two review authors who worked independently and resolved disagreements by discussion. Conclusions were justified, documented, and incorporated into the reporting of results for each outcome. A ‘high’ level of evidence score was obtained according to the GRADE scoring rule after assessing the risks of inconsistency, indirectness, imprecision and publication bias.

Fig. 12
figure 12

Meta-analysis GRADE assessment. Search strategy of PubMed/Medline. PubMed platform. Searched from 1990 to December 12, 2022. #1 (((diabetes mellitus) AND percutaneous transluminal angioplasty)) OR ((diabetes mellitus) AND stent) (798). #2 ((Percutaneous Transluminal Angioplasty) OR (Transluminal Angioplasty) OR (Endoluminal Repair) OR (Angioplasty) OR (Stent) OR (Endovascular Stent Grafting) OR (Stent Grafting) OR (Stents)) AND ((Diabetes mellitus) OR (Diabetes) OR (Diabetic)) AND ((Restenosis) OR (Graft Restenosis) OR (Restenoses)) AND ((random) OR (randomized) OR (randomised)) (235). #3 ((((diabetes mellitus) OR (diabetes)) OR (melituria)) OR (diabetic)) AND (restenosis) (316)

Full size image

Discussion

At present, PTA or stent implantation is the main treatment for cardiovascular stenosis or occlusion; however, restenosis after endovascular treatment is still a challenge [20]. Indeed, as the number of stent placements has risen to an estimate of over 3 million annually worldwide, revascularization procedures have become much more common [21]. However, restenosis after endovascular therapy is a major problem, and studies have shown that 30% to 50% of patients with coronary ischemic disease experience restenosis after endovascular therapy. To date, DM has been recognized as a high-risk factor for cardiovascular events [7, 22]. Epidemiological investigations have shown that patients with concomitant DM and PAD are at high risk for major complications, such as amputation [23]. Technical progress, such as the application of drug-coated balloons or drug-eluting stents [20, 24], has been found to potentially increase patency after endovascular treatment and thus reduce restenosis. The studies we enrolled included a large number of RCTs that involved drug-coated balloons and drug-eluting stents. Although the incidence of restenosis was significantly lower than that of traditional balloons or bare-metal stents, restenosis remained at ahigh rate and was difficult to resolve.

A study published in 1999 found that the long-term need for TLR increased with higher classes of in-stent restenosis (ISR) (hazard ratio (HR) = 1.7; P = 0.0380) and with the presence of diabetes (HR = 2.8; P = 0.0003) [25]. Taken together with other evidence, DM is suggested to be a strong determinant of restenosis (neointimal hyperplasia) [26, 27]. Michael Jonas et al. [28] used the insulin resistance model of the Zucker fatty rat and found that insulin-resistant Zucker fatty rats developed a thicker neointima and a narrower lumen area 2 weeks after implantation of an abdominal aortic stent compared with normal Zucker lean rats. Additionally, Manikandan Panchatcharam et al. [29] established a femoral artery guide wire injury model in hyperglycemic mice and confirmed that hyperglycemia had an obvious accelerating effect on intimal regeneration after vascular injury by promoting smooth muscle cell proliferation and migration. These animal studies demonstrated the role of hyperglycemia and ISR in regulating the function of vascular smooth muscle cells (VSMCs) and promoting neointimal hyperplasia. Moreover, advanced glycation end products (AGEs) also play an important role in promoting the progression of diabetic vascular disease. Zhongmin Zhou et al. [30] demonstrated a prominently increased accumulation of AGEs and immunoreactivities of receptor for advanced glycation end products (RAGEs) in response to balloon injury in diabetic compared with nondiabetic rats. Additionally, blockade of RAGE/ligand interaction significantly decreased VSMC proliferation in vitro and bromodeoxyuridine (BrdU)-labeled proliferating VSMCs in vivo, suppressed neointimal formation and increased luminal area in both diabetic and nondiabetic rats. These animal studies demonstrated that the pathophysiological features of diabetes, including ISR, metabolic syndrome, hyperglycemia, and increased AGEs, are involved in neointima after balloon dilation or stent implantation.

The current mainstream view is that DM increases the risk of restenosis [31]. Studies have shown that patients with insulin-dependent DM are at particularly high risk for adverse events after percutaneous coronary intervention (PCI) [27]. The universally accepted hypothesis for this phenomenon is that hyperglycemia induces endothelial dysfunction and a proinflammatory state that promote the production of growth factors and cytokines, leading to extensive neointimal formation and thus contributing to the progression of restenosis [32, 33]. Besides the endothelial cell inflammation hypothesis, researches indicated that endothelial progenitor cells played a significant role in restenosis [34, 35]. Balestrieri ML et al [36] revealed that high glucose concentration decreased the quantity of endothelial progenitor cells via SIRT1 signaling pathway. In addition, the prethrombotic environment of patients with diabetes ultimately increases the risk of restenosis [37]. However, until now, there has been no conclusive, large-scale clinical evidence to support this view. Our meta-analysis involved 20 RCTs from multiple countries and time spans, with up to 31,066 patients. The results of the meta-analysis confirmed for the first time that DM is a high-risk factor for restenosis after endovascular treatment in a human cohort.

Studies [38] have shown that the type of glucose-controlling drug can affect postoperative restenosis in diabetic patients after coronary stent or balloon dilation. Their data, especially regarding metformin and thiazolidinediones, indicate beneficial results compared to insulin and sulfonylurea for restenosis. However, no large trials have been undertaken in which the effect of glucose-lowering agents on restenosis is associated with improved outcomes. Indeed, experts believe that maintaining proper glycemic control is crucial for diabetic patients who have undergone revascularization procedures [21]. In several recent prospective studies, high glycemic levels [39, 40] and insulin resistance [41] have increased restenosis rates after coronary stenting or balloon angioplasty, but these results were based on statistical analysis of clinical phenomena, lacking evidence from high-quality controlled studies. Marfella R et al [42] conducted a RCT involving 165 patients with high blood glycemic and ST-segment elevation myocardial infarction (STEMI) undergoing PCI treatment, randomly assigning these patients to an interventional-glycemic-control group and an intensive-glycemic-control group. The results showed that the restenosis rate of patients in the intensive-glycemic-control group after PCI reduced by half (48% and 24%) at 6 months. The study confirmed the positive correlation between blood glucose levels and restenosis. However, it solely concentrated on blood glucose levels without addressing whether the patients had diabetes or specific subtypes. Our research indicates that restenosis of the diabetic patients after interventional treatment is not directly related to blood glucose level. This implies that diabetes may pose other risks aside from high blood glucose, such as AGEs and insulin resistance, which could potentially contribute to restenosis. Although Mone P et al [43] demonstrated the effect of high glucose on the risk of restenosis in STEMI patients without DM, the restenosis of high glucose-DM group (18.5%) is still higher than high glucose non-DM group (14.0%) at one-year follow-up, which indicates that diabetic patients may be influenced by additional pathogenic factors beyond elevated blood glucose levels. In animal studies, it has been demonstrated that hyperglycemia and insulin resistance lead to intimal hyperplasia after vascular injury in rats [28,29,30, 44, 45], which seems to be of crucial importance in determining exaggerated neointimal hyperplasia after balloon angioplasty in diabetic animals. However, our meta-analysis showed that blood glucose levels in diabetic patients did not affect the incidence of restenosis after endovascular therapy (Fig. 5). In animal experiments, hyperglycemia can contribute to neointimal hyperplasia, which was inconsistent with the results of our meta-analysis, suggesting that there may be other important risk factors involved in restenosis in diabetic patients. A prospective observational study involving 377 participants discovered that postoperative restenosis rates differed among patients with type 2 diabetes and acute myocardial infarction (AMI) based on whether they were prescribed oral sodium/glucose cotransporter 2 (SGLT2) inhibitors [46]. The study confirmed that the administration of SGLT2 inhibitors to type 2 diabetes patients was associated with a reduced frequency of ISR-related events, independent of glycemic control. This research highlights the importance of non-glycemic factors in the reduction of postoperative restenosis among diabetes patients. AGEs and their receptor isoforms seem to have an important contribution to both the pathogenesis and clinical outcome of restenosis. AGEs have been shown to promote carotid intimal regeneration in rats after balloon injury [30], suggesting that AGEs may act as a stimulus for restenosis. Cristiano Spadaccio et al. found that soluble RAGE (sRAGE) levels and total circulating AGEs were positively correlated with an increased risk of stent restenosis [47, 48]. This suggests that AGEs may play a more important role in restenosis than hyperglycemia. Currently, there is a lack of clinical studies on the relationship between restenosis and AGE levels, and further RCT studies may be of great significance.

DAPT has become essential in daily clinical practice. In fact, current practice guidelines recommend aspirin and clopidogrel DAPT for patients suffering from CAD or monotherapy for patients with symptomatic PAD, regardless of clinical background [49]. However, there is controversy over the duration of antiplatelet therapy in view of the different lesion sites and stent types [50]. Cristian A Dámazo-Escobedo et al [51]. confirmed that long-term antiplatelet therapy after coronary stenting would be justified by the high incidence of thrombosis-restenosis through a prospective observational study. However, based on the current European guidelines for management after coronary stent placement, there is no consensus on the ideal duration of DAPT to prevent stent thrombosis-restenosis without a significant increase in bleeding risk. In our meta-analysis, based on the circumstances of antiplatelet therapy included in the study after revascularization, we took 6 months of DAPT as the line of demarcation and found that different durations of antiplatelet therapy had no significant effect on the incidence of restenosis after endovascular therapy. Moreover, anticoagulation therapy after revascularization in PAD is particularly important compared to that after coronary revascularization. The COMPASS and VOYAGER PAD RCTs showed that a low-dose oral anticoagulant combined with aspirin (Rivaroxaban 2.5 mg twice a day; Aspirin 100 mg once a day) improved the long-term patency rate of lower extremity artery disease after endovascular therapy, reduced the incidence of major limb adverse events and cardiovascular events, and did not increase the risk of fatal major bleeding [52,53,54]. Regrettably, due to the lack of available anticoagulant therapy regimens in the included studies, we did not perform a subgroup meta-analysis of anticoagulant therapy and restenosis in this study. Anticoagulation combined with antiplatelet therapy may be beneficial in future RCTs.

It is worth mentioning that according to the subgroup analysis, patients with diabetes mellitus had a higher TLR rate, while patients with restenosis detected by angiographic follow-up showed no significant difference. This implies that patients with diabetes may be more prone to symptomatic restenosis and require surgical reintervention, suggesting to clinicians that patients with diabetes may have increased reoperation rates. Therefore, diabetic patients should take this characteristic into full consideration when choosing the type of balloon or stent, such as the choice of a drug-eluting balloon to reduce restenosis. Since this study did not involve comparisons of balloon types or stent types, research in this direction may be of great significance for diabetic patients.

The findings of the meta-analysis involving 31,066 individuals affirm that patients with diabetes mellitus (DM) have a higher risk of restenosis following intravascular treatment. Nevertheless, there are still limitations to this meta-analysis. More detailed baseline characteristics of patients were not obtained from some of the included studies, even after contact these corresponding authors, such as the medical history time of DM, detailed blood glucose levels and hypoglycemic methods (such as diet, exercise or drug therapy), which may be important factors and could affect the analysis of restenosis. Due to the high rank of the GRADE results, we also suggest a cautious interpretation for this meta-analysis, and further high-quality RCTs are needed to improve the current conclusion.

Conclusions

In summary, the findings of this systematic review and meta-analysis provided convincing evidence that patients with DM had an increased risk of primary restenosis after PTA or stenting, suggesting that DM is a high-risk factor for restenosis after endovascular treatment, irrespective of blood glucose level, antiplatelet therapy duration, targeted lesion vessel and continent. In conclusion, our meta-analysis provides a reliable suggestion for the health management of diabetic patients with vascular occlusive disease after endovascular therapy.