Revascularization for Stable Chronic Total Occlusion is Essential for Diabetic Patient with Heart Failure

Background Revascularization is the recommended treatment strategy for patients with heart failure (HF) and coronary artery disease (CAD). However, chronic total occlusion (CTO) is less attempt. Furthermore, there were conicting debates on if diabetic HF patients gained benets from revascularization. As to CTO revascularization, no study answered if it offered benets to diabetic HF patients. Methods Diabetic patients with stable CTO and HF were consecutively enrolled in this retrospective cohort study. Based on treatment strategies to the CTO vessel, patients were assigned to successful revascularization (CTO-SR) or medical therapy (CTO-MT) group. The primary endpoints were major adverse cardiac events (MACE). Subgroup analysis were performed based on left ventricular ejection fraction (LVEF) and relevant baseline variables. Results A population of 680 patients were enrolled in the present study: 344 patients in the CTO-MT group, and 336 patients in the CTO-SR group. After a median follow-up of 34 months, CTO-SR was superior to CTO-MT in MACE (adjusted hazard ratio [HR]: 0.462, 95% conference interval [CI]: 0.337-0.634), which could mainly due to the superiority in cardiac-death and TVR. Propensity matching analysis also conrmed CTO-SR’s superiority (HR: 0.494 [0.337-0.725]). Subgroup analysis further conrmed a consistent superiority in patients with LVEF ≥ 40%, but not in those with LVEF<40%. Conclusions For patients with diabetes, HF and stable CTO, CTO-SR was superior to CTO-MT. CTO-SR’s superiority was consistent in patients with LVEF ≥ 40%, but not for patients with LVEF<40%. Trial registration This study was not registered in an open access database.

high-risk features. However, Professor Galassi [18] found that even for CTO patients with LVEF 35%, CTO-PCI could offer patients LVEF improvement and midterm clinical bene ts. Thus, the risk/bene t ratio is favorable for this high-risk procedure. Urgent evidences are needed to answer if diabetic patients with HF need an invasive CTO revascularization procedure. By now, no study directly focused on this.
Thus, we designed the present study which enrolled stable CTO patients with diabetes and HF. We try to answer if successful CTO revascularization (CTO-SR) offered patients clinical bene ts when comparing with conservative medicine therapy (CTO-MT).

Study design and population
Between January 2007 to December 2017, 2502 diabetic patients with one CTO were enrolled in an allcomer, retrospective cohort study. This present study extracted only patients with heart failure. Inclusion Of note, CTO was de ned as total occlusion of the native coronary artery (Thrombolysis In Myocardial Infarction [TIMI] ow grade 0) with a duration of at least 3 months. Occlusion duration was calculated from previous angiography or rst episode of angina. For those who express no symptoms, we arti cially consider them meeting our criteria.
Grouping CABG, PCI and MT were all alternative treatment methods for CTOs, and the nal decision was based on both physician's and patient's choice. Patients were assigned to different groups based on their nal treatment strategy to the CTO: successful revascularization (CTO-SR) or medical therapy (CTO-MT). To sum up, CTO-SR included patients with CTO-CABG and successful CTO-PCI. CTO-MT contained patients with failed CTO-PCI, failed CTO-CABG and initial CTO-MT. Successful CTO-PCI was de ned as a residual stenosis of <20% and a TIMI ow ≥2 [19] after drug eluting stents implantation. Failed CTO-CABG was de ned as grafting to other arteries but not the CTO.

Procedure
Baseline information was extracted from a hospital information system (HIS) by independent investigators who were trained beforehand to ensure consistent. Basic and procedural cine angiograms were all re-reviewed by two experienced interventional physicians using standard de nition.
We performed a minimum of 12 months follow-up procedure (around January 2019) by phone call. For patients with records of re-hospitalization in Beijing Anzhen Hospital, interesting information was also extracted from the HIS. We set up an adjudication board (Shuzheng Lyv, Hong Liu, Quanming Zhao, and Fei Yuan) who were blind to patients groups. Certi cations of clinical endpoints were based on the board.
For all patients, an optimal medical therapy is recommended. Of note, antiplatelet drugs with aspirin at 100mg/day, and/or clopidogrel at a dose of 75mg/day or ticagrelor at 90 mg twice daily. High dose statin or/and ezetimibe to reduce the level of low-density lipoprotein cholesterol was needed. Other therapies, like antianginal therapy, angiotensin-converting enzyme inhibitor, diuretics, digitalis glycosides, aldosterone receptor antagonist, cardiac resynchronization therapy, et al. were recommended in needed patients.
De nition and endpoints HF [3] was de ned as the presence of dyspnea or equivalent symptoms due to cardiac causes. When gathering the information of heart failure, we extracted the diagnosis, NYHA grade, patients' symptom and LVEF. The certi cations of HF were also based on the adjudication board.
The de nition of clinical endpoints were following the Academic Research Consortium (ARC) [20] and the third universal de nition of MI [21]. The primary clinical endpoint was prede ned as major adverse cardiac events which was a composite of cardiac-death, target vessel revascularization (TVR) and repeat nonfatal HF. Cardiac-death was de ned as death of cardiac, unwitnessed or unknow cause. TVR was de ned as repeat revascularization to the previous CTO vessel. Only those with a record of rehospitalization due to heart failure was considered as repeat heart failure. Other clinical endpoints included all-cause death, repeat revascularization and repeat nonfatal myocardial infarction (MI). allcause death was de ned as death of any cause. Repeat revascularization was de ned as revascularization to both previous CTO and non-CTO vessels. Repeat MI was de ned as an overall consideration of angina or equivalent symptoms, electrocardiogram variation and changes in myocardial damage biomarkers. We used only repeat nonfatal MI and repeat nonfatal HF as our endpoint.
Diabetes was de ned as a previous diagnosis or a new diagnosis (a fasting blood glucose level ≥7.0 mmol/L or a glucose level two hours after a meal of ≥11.1 mmol/L detected on more than 2 occasions)

Statistic analysis
Continuous variables are presented as means±SDs (normal distribution) or medians with interquartile ranges (skewed distribution). Comparisons were calculated by using the Mann-Whitney U test or the Student's t test, where appropriate. Categorical items are presented as counts (percentages) and were analysed by using the chi-square test or the Fisher exact text.
For clinical endpoints, the Kaplan-Meier method were used to draw survival curves and comparisons were performed by using the log-rank tests. The unadjusted hazard ratios (HRs) were analysed by using the univariate Cox proportional hazard regression model. We also performed a multivariate Cox regression. Potential adjusted variables were selected based on both the univariate Cox regression (by evaluating all variables in Table 1 and applying a threshold of P≤0.1) and the relevant clinical implications. The adjusted hazard ratio (HR) was calculated based on: age, chronic obstructive pulmonary disease (COPD), peripheral vascular disease (PVD), prior MI, hyperuricemia, left ventricular end diastolic diameter (LVEDD), ventricular aneurysm, reginal wall motion abnormality (RWMA), single vessel disease, left anterior descending CTO (LAD-CTO) and Rentrop grade≥2.
A propensity score-matched analysis was also performed. All variables listed in Table 1 (except for the retrograde approach) was included in the model. A 1:1 matching was performed by applying the nearestneighbour method (calliper value=0.02). Absolute standardised differences (ASDs) were used to evaluate the effect of matching. ASDs <10.0% showed a relatively good balance. Baseline variables and clinical endpoints after propensity matching were re-compared.
Subgroup analysis was performed based on LVEF in all clinical endpoints.
Statistical analyses were performed by utilizing SPSS 24.0 (SPSS Inc., Chicago, Illinois, USA) and Stata 14.0 (Stata, College Station, TX, USA). A P-value ≤0.05 was deemed to be statistically signi cant.

Baseline characteristics
Patients enrollment was shown as Figure 1. To reiterate, 680 HF patients were selected from 2502 patients: 344 patients in the CTO-MT group, and 336 patients in the CTO-SR group. Patients treated with CTO-SR were younger, more likely to be male, and had a higher prevalence of smoking, aspirin uptake, LAD-CTO and good collateral circulation (Rentrop grade≥2). While patients treated with CTO-MT had a higher prevalence of Prior PCI, NYHA class , LCX-CTO, nitrites uptake, stain uptake and P2Y12 inhibitor uptake (Table 1).

Procedural complications and in-hospital adverse events
We summarized procedural complications in patients treated with PCI. 11 patients who were treated with failed CTO-PCI (CTO-MT group) suffered coronary artery dissection. There is no contrast extravasation, pericardial effusion and emergency surgery. While for patients treated with successful CTO-PCI (CTO-SR group), ve patients suffered contrast retention. There was also no other complications occurred.
In-hospital adverse events (mortality) occurred in ve patients. All were treated with CTO-CABG (CTO-SR group). We excluded these ve patients when analyzing long-term clinical outcomes.
Another subgroup analysis showed that the superiority of SR did not differ in all subgroups ( Figure 5).

Main ndings
To our knowledge, this is the rst study which focused on diabetic CTO patients with heart failure and try to secure which treatment strategy will bene t patients. We observed that: (1) CTO-SR bene ts patients with lower risk of MACE than CTO-MT. The bene t was mainly attribute to a lower incidence of TVR and cardiac-death. (2) CTO-SR's superiority was consistent in patients with 40%≤LVEF<50% and LVEF≥50%, But not in those with LVEF<40%. (3) Patients who were treated with CTO-MT were associated with more severe situations: older and poor collateral circulation. However, age, collateral circulation, RWMA or ventricular aneurysm did not in uence the superiority of CTO-SR. successful CTO-PCI was associated with promising clinical outcomes, as well as LVEF improvements. However, they compared successful CTO-PCI with failed CTO-PCI, and diabetes contains only 30% of enrolled patients. Furthermore, they contained only patients with viable myocardium. Thus, their results could not expand to general diabetic CTO patients with HF. In the present study, we demonstrated that diabetic patients with CTO and HF, successful CTO revascularization could offer patients survival bene ts and other clinical bene ts, which will surely promote the process of revascularization attempt, regardless of the high-risk feathers.
In the present study, we failed to see a consistent superiority of CTO-SR in patients with LVEF<40%, which is opposite to Professor Galassi [18] and Professor Wolff [15]. We assumed the following possibilities: (1) Sample size enrolled in subgroups of LVEF<40% were small. Even though 30 MACE occurred in CTO-MT versus 13 MACE occurred in CTO-SR, after analysis with Cox regression, no statistical difference was observed. Thus, further study with larger sample size was needed. (2) Detection of viable myocardium. Professor Galassi [18] enrolled only patients with viable myocardium and reported positive outcomes of successful CTO-PCI. However, in the present study, viable myocardium was not routinely detected. It might be assumed that only patients with viable myocardium can get bene t from revascularization. Yet Professor Wolff and colleagues [15] reported a signi cant survival improvement for patients with LVEF<40% who were treated with CABG, in spite of myocardium viability. Furthermore, in the present study, we also observed clinical bene ts for patients with LVEF 40%, regardless of myocardium viability.
Thus, a reasonable assumption might be for patients with diabetes, CTO and LVEF<40%, detection of myocardium viability was necessary. Further studies were required to con rm this.
On the other hand, our outcomes in the subgroup of LVEF<40% was consistent with STICH trial [16] which showed that CABG did not offer diabetic patients clinical bene ts. We thought diabetic status might in uenced the results. Diabetes is reported to promote the process of coronary atherosclerosis and associated with poor clinical outcomes [29-31], which surely contributed the poor therapeutic effect of CTO-SR in patients with LVEF<40%.

Is viable myocardium indispensable in diabetic patients with HF and CTO?
The ESC/EACTS guidelines recommended positron emission tomography (PET) may be considered for the evaluation of myocardial viability in HF patients, but with a low recommended level: IIb [13]. Furthermore, the correlation between viable myocardium and revascularization bene ts was uncertain [13,32]. Additionally, the detection value of PET might be limited in diabetic patients due to insulin resistance [33,34]. Thus, it is unknown which kind of patients could get certain bene ts from viability testing. In the present study, we observed both patients with RWMA and aneurysm got bene ts from revascularization. However, patients with LVEF<40% failed to gain bene ts. Therefore, it is safe to state that viable myocardium should be detected for diabetic CTO patients with lower LVEF.

The effect of collateral circulation on clinical outcomes
In the present study, poor collateral circulation was more common in patients treated with CTO-MT.
However, collateral circulation did not in uence the superiority of CTO-SR.
Collateral circulation provided blood irrigation for CTO territories. Some argued that for patients with welldeveloped collaterals, revascularization is dispensable. However, collaterals are not su cient to provide enough blood supply for CTO territory. Furthermore, coronary steel [35] and lower blood pressure [36] can also decrease the protective in uence of collaterals. Thus, revascularization should be performed. In the present study, we demonstrated that even for patients with well-developed collateral circulations, successful revascularization offered patients low incidence of MACE and other clinical bene ts, which is consistent to previous studies [37] and offered positive evidence on revascularization attempt.

Limitation
This present study was restricted by the following items: (1), This study was a single-center, retrospective cohort study, though we performed propensity matched analysis and multivariable analysis, we could not balance all potential confounders. (2), We did not detect viable myocardium. (3), in the present study, the validation of repeat nonfatal MI, repeat nonfatal HF and TVR was di cult. To assure accuracy, only individuals who had records of re-hospitalization in our Hospital or who provided a written proof of diagnosis (by WeChat) were identi ed as meeting de nitions of these endpoints, which surely brought about some bias. (4), We only enrolled diabetic patients with CTO, which indicated that our results could not be expended to the entire CTO population.

Conclusions
In conclusion, for patients with diabetes, HF and stable CTO, CTO-SR was superior to CTO-MT. CTO-SR's superiority was consistent in patients with LVEF≥50%, but not for patients with LVEF<40%.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and analyzed for this current study are available from the corresponding author upon reasonable request.

Competing interests
All authors declare that they have no competing interests.

Funding
The present work was supported by the Ministry of Science and Technology of the People's Republic of China, State Science and Technology Support Program (No.2011BAI11B05) and the Beijing Lab for Cardiovascular Precision Medicine, Beijing, China (PXM2019_014226_000023). The project director of all projects was professor Shuzheng Lyu (the corresponding author) and the funds were provided by the government or public institution to support medical research. All these institutions had no effect on the research process.     HR: hazard ratio; CI: conference interval; TVR: target vessel revascularization; other abbreviations as in Table  S1.