Successful revascularization versus medical therapy in diabetic patients with stable right coronary artery chronic total occlusion: a retrospective cohort study

The present study was a retrospective cohort study that was performed at Beijing Anzhen Hospital (Beijing, China). From January 2007 to December 2017, a population of 2502 stable patients with one main stem CTO were selected. In this study, we enrolled patients who had only RCA-CTO. The inclusion criteria were as follows: (1) diabetic patients with only one main stem CTO that was diagnosed by angiography and (2) patients with RCA-CTO (American Heart Association segment maps 1, 2 and 3) who had stable clinical presentations of symptomatic stable angina or silent ischemia. We excluded those patients with (1) a prior CABG history; (2) left main coronary artery disease (stenoses ≥ 50%); (3) a history of myocardial infarctions (MIs) due to a non-CTO artery within 30 days; and (4) patients with cancer or other diseases that may have confounded the endpoints.

Patients were assigned to the medical therapy (CTO-MT) group or to the successful revascularization (CTO-SR) group according to their final treatment strategy of the CTO vessel. Of note, the CTO-SR group enrolled patients who were treated with CTO-CABG or with successful CTO-PCI. The CTO-MT group enrolled patients who were managed by failed CTO-CABG, failed CTO-PCI or initial medical therapy of the CTO vessel. The definition of successful CTO-PCI was obtaining a residual stenosis of < 20% and a thrombolysis in myocardial infarction (TIMI) flow ≥ 2 [21] after implantation of a drug eluting stent to the CTO vessel. Failed CTO-CABG was defined as performing grafts to other arteries but not the RCA territory.

An optimal medical treatment was recommended to both groups of patients who were managed by SR and MT. To reiterate, an antiplatelet therapy involving the use of aspirin and/or clopidogrel (ticagrelor), a lipid lowering therapy, an antianginal therapy and other secondary prevention therapies were all recommended to be prescribed for the patients as needed.

The strategies of MT or revascularization (CABG or PCI) were both alternative methods for treating a CTO vessel, and the final strategy decision was made based on the preferences of both the physicians and the patients. During the PCI procedure, the uses of contemporary techniques, such as microcatheters, bilateral injections and retrograde approaches, among others, were left to the discretion of the operator.

The definition of RCA-CTO was a TIMI flow grade of 0 within the RCA vessel with a duration of at least 3 months [22]. The occlusion interval was calculated from the last episode of myocardial infarction, from the first episode of stable angina or from a previous angiogram. For those patients who exhibited no clinical symptoms, we artificially identified them as meeting our inclusion criteria. Diabetes mellitus was defined as a previous diagnosis of diabetes before hospitalization or a new diagnosis with either fasting blood glucose levels ≥ 7.0 mmol/L or a glucose level at 2 h after a meal of ≥ 11.1 mmol/L on more than two occasions [23]. A prior MI was confirmed according to the electrocardiogram, the ultrasonic cardiogram or the records from a hospital information system. An angiographic stenosis > 50% was considered to indicated a diseased vessel. Patients with a CTO vessel that represented the only diseased vessel were diagnosed with single-vessel disease. Systolic heart failure consisted of only heart failure with a reduced ejection fraction (HFrEF) and heart failure with a mid-range ejection fraction, which was defined as a left ventricular ejection fraction (LVEF) < 50% with either the presence of dyspnea or equivalent symptoms, while diastolic heart failure exhibited a preserved ejection fraction (HFpEF) [24]. Both right dominance and codominance were considered to be right dominance.

We predefined the primary clinical endpoint as all-cause death, which was mortality due to any reason. Other clinical outcomes included cardiac death, repeat nonfatal MI, repeat revascularization and target vessel revascularization (TVR). Cardiac death was defined as the following by the Academic Research Consortium (ARC) [25]: any death that is related to a cardiac reason, an unwitnessed mortality or an unknown cause of mortality. Noncardiac death was defined as death from a certain noncardiac cause. Probable/definite cardiac death was defined as death of certain (myocardial infarction, heart failure, etc.) or probable (sudden death) causes. The definition of a repeat MI was determined by following the third universal definition of MI [26]: a complicated consideration of ischemic symptoms, electrocardiograms and myocardial damage biomarkers. In this study, we collected data from only repeat nonfatal MI patients. The definition of repeat revascularization was an unplanned revascularization to the target vessel (RCA) or to the nontarget vessel. TVR was defined as any surgical bypass or percutaneous intervention of the target vessel (RCA).

The data collection was managed by experienced raters who were trained beforehand in order to ensure accordance. The following items were selected for analysis: age, gender, clinical history (hypertension, dyslipidaemia, prior MI and systolic heart failure, among others), inspection information, medical data and angiogram information. For the angiogram data, experienced, interventional physicians rescanned the cine angiograms by using standard definitions in order to reduce bias.

A minimum follow-up period of 12 months was predefined. A pre-designed chart including all of the follow-up items was applied. The follow-up procedure was performed by experienced investigators who were blind to the patients’ assigned groups. A phone call was the preferred follow-up method. For patients who had records of re-hospitalization in the Beijing Anzhen Hospital, noteworthy information was also obtained from the hospital information system. In the present study, we attempted to minimise the crossover between the groups. Thus, for patients who changed their treatment strategy during the follow-up, we ended the follow-up process at the moment of the treatment change. The follow-up information was evaluated by an adjudication board (Shuzheng Lyv, Hong Liu and Fei Yuan) who were blind to the patients’ assigned groups. The certifications of the clinical endpoints were also based on the decisions of the board.

The continuous variables are presented as the means ± SDs (normal distribution) or medians with interquartile ranges (skewed distribution). Comparisons were performed using Student’s t-test or the Mann–Whitney U test, where appropriate. The categorical items are presented as numbers and percentages and were analyzed using the Chi-square test or Fisher’s exact text. Events per 1000 patient-years and survival curves using the Kaplan–Meier method were applied for all of the endpoints. A univariate Cox proportional hazard regression model was used to calculate the unadjusted hazard ratios (HRs). After the univariable analysis, a multivariate Cox regression model was performed. The potential adjusted factors were selected according to both the univariate Cox regression model (via the analysis of all of the variables listed in Table 1 and by using a threshold of P ≤ 0.2) and the relevant clinical implications. The adjusted HR (model 1) was calculated by utilizing the variables that exhibited statistical significance, as demonstrated by the univariate Cox regression: age, chronic kidney disease (CKD), chronic obstructive pulmonary disease (COPD)/asthma, prior MI, systolic heart failure (HF), LVEF, regional wall motion abnormality (RWMA), single-vessel disease, triple-vessel disease and syntax scores. Afterwards, we added further potential clinically relevant factors to calculate the adjusted HR (model 2): sex, peripheral vascular disease (PVD) and HbA1c. To further evaluate the difference between the CTO-SR and CTO-MT groups, we performed a propensity score-matched analysis. First, a propensity score utilizing a logistic regression model, which absorbed all of the variables listed in Table 1 (except for the retrograde approach), was calculated. The pairs were then matched at a 1:1 ratio by utilizing a nearest-neighbor matching method (caliper value = 0.02). We assessed the balance of the variables through absolute standardized differences (ASDs). ASDs < 10.0% revealed a relatively small imbalance. The baseline items of the propensity-matched population were then reanalyzed using the same methods as previously described. The clinical endpoints were also reanalyzed with the Kaplan–Meier method and the log-rank tests. The univariate Cox proportional hazard regression model was applied to calculate the HRs. To evaluate whether right coronary dominance affects clinical endpoints, an additional analysis of patients with only right dominance was also performed between the CTO-SR and CTO-MT groups.

Subgroup analysis was performed between subgroup populations: successful CTO-PCI versus initial CTO-MT, successful CTO-PCI versus failed CTO-PCI, successful CTO-PCI versus CTO-CABG, CTO-CABG versus initial CTO-MT and CTO-CABG versus failed CTO-PCI.

Another subgroup analysis, including sex (male/female), age (< 60 years old/≥ 60 years old), prior MI (yes/no), RWMA (yes/no), HbA1c (< 7.0%/≥ 7.0%), single-vessel disease (yes/no), systolic heart failure (yes/no), Rentrop grade ≥ 2 (yes/no) and syntax score (< 22/≥ 22), was also conducted using a multivariate Cox regression model. The covariates that were applied in this Cox regression model were the same as the items applied in model 2. Of note, we only performed this subgroup analysis for all-cause death and cardiac death.

Statistical analyses were performed using SPSS 24.0 (SPSS Inc., Chicago, Illinois, USA) and Stata 14.0 (Stata, College Station, TX, USA). A two-tailed P-value ≤ 0.05 was considered statistically significant.

From January 2007 to December 2017, a total of 943 stable RCA-CTO patients with diabetes were consecutively enrolled in the present study (Fig. 1). Of these patients, 443 (46.98%) were managed by MT (the CTO-MT group), and 500 (53.02%) were managed by successful revascularization (the CTO-SR group). The CTO-MT group enrolled patients with initial CTO-MT (n = 233), failed CTO-PCI (n = 191) and failed CTO-CABG (n = 19). The CTO-SR group enrolled patients with successful CTO-PCI (n = 309) and CTO-CABG (n = 191).

The baseline clinical and angiographic characteristics are listed in Table 1. To reiterate, compared to those in the CTO-SR group, patients who were treated with MT had a higher prevalence of dyslipidemia, hyperuricemia, prior PCI, prior stroke, statin intake, nitrite intake and single-vessel disease, whereas these patients were less likely to be male and had a lower prevalence of right dominance coronary circulation and good collateral circulation (Rentrop grade ≥ 2). Patients who were treated with SR were more likely to undergo the retrograde approach. For the other variables, there were no significant differences between the groups.

In-hospital mortality occurred in four patients, and these patients were all treated with CABG: three patients in the CTO-SR group and one patient in the CTO-MT group. When analyzing the mid-term clinical endpoints, we excluded these four patients. Procedural complications occurred in seven patients who were treated with failed CTO-PCI (the CTO-MT group): four patients exhibited contrast extravasation (no pericardial tamponade), two patients with coronary artery dissection, and the guidewire broke and remained in the distal iliac artery in one patient. Additionally, complications occurred in seven patients who were managed by successful CTO-PCI: one patient exhibited contrast extravasation (no pericardial tamponade), five patients had contrast retention and one patient experienced sudden cardiac arrest during the procedure.

Table 2 shows the clinical endpoints of this study. After a mid-term follow-up (CTO-MT: median: 42.00 months, interquartile range [IQR]: 24.00–78.25 months; CTO-SR: 48 [29.00–90.50] months), the data on all-cause death was successfully obtained from 874 (93.08%) patients: 28.15 all cause death per 1000 patient-years in the CTO-MT group versus 12.17 all cause death per 1000 patient-years in the CTO-SR group (P < 0.001). The univariate analysis (unadjusted HR: 0.423, 95% conference interval [CI] 0.270–0.663) demonstrated that CTO-SR was superior to CTO-MT for all-cause death. This superiority was also confirmed by the multivariate analysis (adjusted HR [model 1]: 0.429, 95% CI 0.269–0.682; adjusted HR [model 2]: 0.445, 95% CI 0.278–0.714).

When considering the outcome of cardiac death, CTO-SR also demonstrated its superiority in events per 1000 patient-years (22.23 versus 6.90, P < 0.001), the univariate (unadjusted HR: 0.311, 95% CI 0.177–0.547) and multivariate (adjusted HR [model 1]: 0.308, 95% CI 0.172–0.550; adjusted HR [model 2]: 0.307, 95% CI 0.169–0.557) models. So as to probable/definite cardiac death (unadjusted HR: 0.297, 95% CI 0.137–0.644; adjusted HR [model 1]: 0.309, 95% CI 0.138–0.693; adjusted HR [model 2]: 0.341, 95% CI 0.151–0.771). The superiority of patients who were managed by SR was consistent for the clinical endpoints of repeat nonfatal MI (unadjusted HR: 0.493, 95% CI 0.266–0.915; adjusted HR [model 1]: 0.466, 95% CI 0.245–0.888; adjusted HR [model 2]: 0.513, 95% CI 0.264–0.998), repeat revascularization (unadjusted HR: 0.683, 95% CI 0.509–0.915; adjusted HR [model 1]: 0.689, 95% CI 0.512–0.928; adjusted HR [model 2]: 0.702, 95% CI 0.516–0.955) and TVR (unadjusted HR: 0.591, 95% CI 0.406–0.862; adjusted HR [model 1]: 0.587, 95% CI 0.401–0.858; adjusted HR [model 2]: 0.613, 95% CI 0.413–0.909) (Table 2 and Fig. 2).

After a 1:1 propensity matching, 286 patients were included in both groups. The ASDs after matching were all less than 10.0%, except for drinking (see Additional file 1: Figure S1). None of the baseline variables showed significant differences (see Additional file 1: Table S1), thus indicating a good matching balance. For the clinical endpoints, the superiority of CTO-SR was also observed for all-cause death (HR: 0.502, 95% CI 0.291–0.865, P = 0.013), cardiac death (HR: 0.331, 95% CI 0.161–0.681, P = 0.003), repeat revascularization (HR: 0.639, 95% CI 0.440–0.927, P = 0.018) and TVR (HR: 0.574, 95% CI 0.355–0.930, P = 0.024) when compared with CTO-MT (Table 3 and Fig. 3). Regarding repeat nonfatal MI, the trend was also consistent (HR: 0.571, 95% CI 0.277–1.177, P = 0.129).

To minimize the influence of the dominance artery, an additional analysis that included only patients with RAC dominance was performed. We found that the superiority of CTO-SR was consistent in all clinical endpoints as in the whole population (see Additional file 1: Table S3, Fig. 4). The baseline characteristics are also provided in Additional file 1: Table S2.

Another subgroup analysis that compared successful CTO-PCI with failed CTO-PCI, initial CTO-MT and CTO-CABG, found that successful CTO-PCI provided patients with a benefit in terms of mortality when compared with initial CTO-MT (all-cause death: adjusted HR 0.498 [0.251–0.987]) and failed CTO-PCI (all-cause death: adjusted HR 0.315 [0.154–0.642]). This benefit can mainly be attributed to cardiac causes (cardiac death: 0.371 [0.154–0.892] and 0.195 [0.082–0.465]; probable/definite cardiac death: 0.463 [0.126–1.701] and 0.151 [0.041–0.563], respectively) (see Additional file 1: Tables S4–S7, Fig. 4). Regarding successful CTO-PCI versus CTO-CABG, CTO-CABG gained more benefit from repeat revascularization (adjusted HR 4.459 [2.044–9.726]) and TVR (21.676 [2.896–162.24]) (see Additional file 1: Tables S8, S9, Fig. 4).

We also compared CTO-CABG with both initial CTO-MT and failed CTO-PCI. Although statistical significance was not observed in all endpoints, CTO-CABG was superior to both subgroups in all-cause death, cardiac death, repeat revascularization and TVR. (see Additional file 1: Table S10–S13, Fig. 4).

Subgroup analysis based on clinically relevant variables showed that the superiority of CTO-SR in all-cause death was consistent in all subgroups, except for patients with single-vessel disease (HR: 1.245 [0.206–7.527]). The same results were also observed in cardiac death (Fig. 5). However, only 8 all-cause deaths and 4 cardiac deaths occurred in patients with single-vessel disease. Furthermore, no significant difference was observed in patients with/without single-vessel disease (P > 0.05).

By multivariate analysis (Fig. 6), CTO-SR was a protected predictor of all-cause death (HR: 0.445, 95% CI 0.278–0.714) and cardiac death (HR: 0.307, 95% CI 0.169–0.557). Conversely, increases in syntax score (per 1 score, HR: 1.097, 95% CI 1.043–1.154) and age (per 1 year, HR: 1.044, 95% CI 1.020–1.068) predicted a worse probability for all-cause death. Additionally, an increase in the syntax score (per 1 score, HR: 1.104, 95% CI 1.043–1.169) was also related to a higher incidence of cardiac death.

To our knowledge, this is the first study that focused on stable RCA-CTO patients with diabetes and that attempted to determine if SR for RCA-CTO offered patients better clinical benefit than that offered by conservative MT. After a mid-term follow-up, our main findings are as follows: (1) CTO-SR offered patients more clinical benefits, which was demonstrated by both the multivariable Cox regression analysis, a propensity score-matched analysis and the subgroup analyses. (2) The subgroup analyses revealed that mortality benefits come from both successful CTO-PCI and CTO-CABG. While, the benefit of repeat revascularization and TVR mainly comes from the CTO-CABG subgroup. Furthermore, CTO-CABG also offered patients more repeat revascularization and TVR benefits when compared with successful CTO-PCI. (3) Syntax score was a harmful predictor of cardiac death. However, neither collateral circulation nor syntax score influenced the survival benefits of CTO-SR. So as to glycemic control (HbA1c).

It is well-known that the territory of the RCA is smaller than that of the LAD. Additionally, good collateral circulation has more often been observed in RCA-CTO [27]. Thus, some experts have argued that successful revascularization for RCA-CTO will not benefit patients. Previous studies [5, 11] reported that successful PCI for RCA-CTO did not offer patient survival benefits, which may provide support for this plausible notion. However, Kalandyk and colleagues [28] reported that RCA-CTO increased the incidence of major adverse cardiac cerebrovascular events in patients undergoing CABG, which might mainly due to the smaller number of grafts to the RCA, indirectly indicating that RCA-CTO should be treated with revascularization. Direct positive evidence was also provided by Mitomo and colleagues [6]. However, all of the above studies focused on the entire CTO population. Diabetic status was associated with fast atherosclerosis progression and poor clinical outcomes [18, 29‐33]. Under such a special situation, whether revascularization should be performed is unknown. Our study demonstrated that CTO-SR offers this specific patient group both survival benefits and other clinical benefits (lower incidence of repeat revascularization, TVR and repeat nonfatal MI). Moreover, we found that successful CTO-PCI offers patients only a mortality benefit, while CTO-CABG offers patients more clinical benefits than successful CTO-PCI does. This will surely better direct clinical practice.

Regarding procedural complications, in the present study, we calculated complications mainly from the PCI procedure. The incidence was low: seven for successful CTO-PCI and seven for failed CTO-PCI. After follow-up, we observed only six instances of repeat-revascularization (all were TVR) in those patients (four patients with successful CTO-PCI and two patients with failed CTO-PCI). No death or other adverse events occurred. We thought the influence of procedural complications on mid-term clinical outcomes might be slight.

Previous findings can provide some clues to explain this issue. Prior studies [10, 34] have focused on patients with both RCA-CTO and other severe situations (older age, unprotected left main artery) and reported a higher incidence of mortality or cardiac mortality. A smaller RCA (left coronary dominance) is also a predictor of all-cause death in CTO patients [35]. Thus, the importance of RCA will be more obvious in patients with other severe conditions.

When considering diabetes, the SWISSI study [36] demonstrated that diabetes was a predictor of progressive coronary atherosclerosis disease (OR: 19.01, P = 0.026). The DIABETES study [37] reported that 50% of repeat revascularization procedures were due to progression in another vessel that was not previously treated. This finding was also confirmed by other experts [38]. We speculated that RCA-CTO is different in patients with diabetes because it is an inherently severe situation and exhibits a relatively fast progression, even though the RCA area might be smaller than the LAD area. One study [39] examined diabetic patients with a non-infarct-related CTO who suffered a ST-elevation myocardial infarction and demonstrated that CTO increased the risk of mortality when compared with those without CTO, which partially confirmed our speculation. Moreover, diabetic patients suffered a higher risk of any cardiovascular event than the risk in patients without diabetes [17]. From that perspective, the aim of successful revascularization is not only to recover the blood supply of the RCA area but also to support potential collateral circulation to either the LAD or the left circumflex coronary artery (LCX) in the case of an acute coronary occlusion.

Another possible explanation is electronic stability. The VACTO Primary Study [40] demonstrated that CTO was an independent risk factor for predicting the incidence of ventricular arrhythmias. Perfusion therapy can decrease the incidence of fatal ventricular arrhythmias [41]. Thus, electronic stability might partially explain the improvement in survival after successful revascularization.

To summarize, the inherent severity of the coronary artery situation in diabetic patients may make RCA-CTO different in diabetic patients than in the traditional patient population. The highlight of our study was demonstrating the successful revascularization benefits of patients in terms of not only survival items but also other clinical items. Furthermore, we demonstrated that successful CTO-PCI mainly offers patient survival benefits.

In the present study, we observed that neither collateral circulation nor syntax score affected the superiority of CTO-SR, and neither did HbA1c.

In terms of collateral circulation, our results were consistent with those from a previous study [41], which enrolled CTO patients with well-developed collateral circulation and demonstrated that revascularization reduced the incidence of cardiac death. Mechanically, although good collateral circulation might offer blood to the occluded area of the RCA, the relationship between collaterals and clinical events remains unclear [30, 42, 43]. Additionally, low diastolic blood pressure [44] and coronary steel [45] can also reduce the protective effects of collaterals. Furthermore, FFR studies [46, 47] observed that collateral circulation affected the evaluation of donor vessel FFR and, hence, the optimal treatment strategy, which might lead to adverse cardiac events. Thus, reperfusion of the RCA-CTO will surely reduce myocardial ischemia as well as cardiac events. On the other hand, diabetic patients were often associated with impaired collateral circulation [23], which makes this reperfusion more essential to perform. In summary, our findings will correct the plausible notion that SR and medical therapy are both acceptable treatments for RCA-CTO patients with well-developed collateral circulation.

When considering coronary severity, we demonstrated that syntax score was a predictive risk factor for mortality. However, the syntax score did not influence the protective effect of CTO-SR. Even though for patients with single-vessel disease, CTO-SR might not offer survival benefit. However, only 209 patients were included in this subgroup, and mortality only occurred in eight patients (cardiac death: four patients), which reduced the strength of the evidence. Further studies that focus on only a single-vessel RCA-CTO in diabetic patients are needed to verify our results.

Some experts observed a connection between glycemic control and clinical outcomes [48]. However, in the present study, we observed that the HbA1c level did not predict mortality. We speculated that the reason might be that the CTO-SR procedure is the main beneficial factor for patients with RCA-CTO. Further studies are still needed to confirm the benefits of glycemic control.