Role of diabetes mellitus in acute coronary syndrome patients with heart failure with mid-range ejection fraction underwent percutaneous coronary intervention ： a 3-year case series follow-up retrospective study

Backgroud Data are limited about the effect of diabetes mellitus (DM) on prognosis of acute coronary syndrome (ACS) patients with heart failure with mid-range ejection fraction (HFmrEF) underwent percutaneous coronary intervention (PCI). This study aimed to investigate the relationship between type 2 diabetes mellitus(T2DM) and the 3-year outcomes in such population. Methods 377 ACS patients with HFmrEF (left ventricular EF 40–49%) underwent PCI (132 diabetic and 245 nondiabetic patients) were included into analysis. The primary outcome was a composite endpoint of all-cause death or HF rehospitalization. Cox proportional-hazards regression analysis and Kaplan–Meier test were used to assess the effect of diabetes on the primary outcome. Sensitivity analysis was conducted with propensity score-matching analysis. Results During a follow-up of three years, diabetic patients had a higher incidence rate of the primary outcome than nondiabetic patients (96.1 vs. 44.6 per 1000 patient-years, incidence rate ratio 2.301, 95% confidence interval 1.334–3.969; P=0.002). Multivariate analysis showed that diabetes mellitus was associated with a significant increase in the composite outcome of all-cause death or HF rehospitalization (adjusted hazard ratio 2.080, 95% confidence interval 1.115–3.878, P=0.021). Sensitivity analysis further confirmed that diabetes mellitus was an independent prognostic factor of long-term adverse outcome for ACS patients with HFmrHF who underwent PCI (adjusted hazard ratio 3.792, 95% confidence interval 1.802-7.980 , P ＜ 0.001). with T2DM was significantly associated with worse long-term outcomes.

fraction (HFrEF) [2]. Heart failure with a mid-range ejection fraction (HFmrEF), as a distinctive phenotype, was first defined as left ventricular ejection fraction (LVEF) 40%-49% in 2016 European Society of Cardiology (ESC) HF guidelines [3,4]. Recent observational studies only focused on description of its characteristics and prognosis of HFmrEF, and patients enrolled in these clinical trials covered different aetiology of HF. In view of above-mentioned facts that previous studies have no selectivity on aetiology, our study concentrated on ACS patients and presented a 3-year retrospective analysis to investigate the role of diabetes mellitus (DM) in ACS patients with HFmrEF underwent percutaneous coronary intervention (PCI).

Patients selection
The ACS patients with HFmrEF underwent PCI was retrospectively screened from the Coronary Angiography and Angioplasty Registry Database of Fujian Medical University Union Hospital, registered from January 2014 to June 2017. The inclusion criteria were as followed: 1) Patients was hospitalized for ACS and underwent PCI, survived to discharge during their first admission; 2) Patients met HFmrEF diagnostic criteria according to the 2016 ESC guidelines [4], reviewed by an expert cardiologist during their hospitalization: a) The presence of symptoms and/or signs of HF; b) LVEF ranged from 40 to 49%; c) Elevated levels of BNP＞ 35 and/or NT-proBNP＞125pg/ml; d) Objective evidence of other cardiac functional and structural alterations underly HF; 3) a 3-year follow-up completed at least. The patients were excluded according to the following procedure: 1) echocardiography data or other clinical information no available; 2) echocardiography confirmed left ventricular ejection fraction (LVEF)≥50% or ＜40%; 3) patients lost to follow-up; 4) HF due to non-ischaemic heart disease (such as valvular heart disease, alcoholic cardiomyopathy, etc.). Through the screening process mentioned above, 377 HFmrEF patients were successfully enrolled in this analysis. The flowchart was showed in Figure 1.

Definition of type 2 diabetes mellitus
Patients were considered as having type 2 diabetes mellitus if they had been previously informed of the diagnosis by a physician or were at a status of glucose-lowering therapy, i.e. insulin, oral hypoglycemic agents, diet and exercise. Patients without previously diagnosed T2MD who required initiation of antihyperglycemic therapy during hospital stay were also considered to have diabetes [5].

Data collection
Patient information and coronary angiography procedural details were collected by an independent trained reviewers blinded to group assignment from hospital databases and recorded in a computerized database. The baseline data include (i) demography; (ii) clinical status; (iii) complications; (iv) comorbidities; (v) electrocardiographic findings; (vi) angiographic and intervention status; (vii) treatment including discharge medications. These patients were divided into diabetic group and nondiabetic group at discharge from the index hospitalization.

Follow-up and Endpoints
All patients were followed up and clinical endpoint events happened within the first 3-year was registered and recorded after the index admission, and follow-up data was collected from the review of hospital charts, or discharge summary review, clinical visits or telephone interviews, which were conducted by independent trained reviewers. The primary outcome was a composite of all-cause death or rehospitalization for HF. Secondary endpoints were a composite of cardiovascular death or rehospitalization for HF, all-cause death, cardiovascular (CV) death, rehospitalization for HF and unplanned revascularization [6]. Cardiovascular death was defined as death due to heart failure, myocardial infarction, cardiogenic shock, sudden cardiac death (SCD), death related to stroke or other cardiac causes. Unexplained death was also regarded as cardiovascular in origin unless obvious non-cardiovascular causes could be identified. Rehospitalization for HF was defined as an admission for decompensated HF after discharge from the index hospitalization, decompensated HF was defined on the basis of symptoms and signs, such as dyspnea, rales, and ankle edema, and the need for intravenous drug therapy, hemodialysis, mechanical circulatory or respiratory support [ 7 ]. Unplanned revascularization was repeat PCI or coronary artery bypass grafting of any vessels excluding staged PCI [8].

Statistical analysis
Continuous variables were presented as mean ± standard deviation or median with interquartile range (IQR), and differences were assessed by Student's t-test, or Wilcoxon rank sum test. Categorical variables were described as percentage (%), and differences were analyzed by Pearson χ2 or Fisher exact test. For analysis of associations between diabetes mellitus and outcomes, incidence rates per 1000 patient-years and incidence rate ratio were calculated for each outcome. Cumulative incidence of outcomes was estimated by using regression estimates from a Cox proportional hazards model including covariates that were either statistically significant(P<0.20) on univariate analysis or clinically relevant. Adjusted hazard ratios (HRs) were estimated by Cox regression model and were presented with 95% confidence interval (CI). The following covariates were taken into adjusted hazard ratios (HRs) For the purpose of the sensitivity analysis, a propensity score was estimated by fitting a logistic-regression model that adjusted for age, gender, Killip functional class (III-IV vs. I-II), hypertension, AF, coronary disease status (multivessel vs. single vessel disease), complete revascularization, ACEI or ARB, calcium channel blocker (CCB). 1:1 pair matching between the two groups was performed by nearest neighbor matching without replacement. The same analysis was performed for the propensity matched cohorts. All tests were two tailed, and P < 0.05 was considered to be statistically significant. All analyses were performed with the SPSS statistical software (version 23.0, IBM, Chicago, USA).          Because of baseline characteristics of patients was disequilibrium, so we used 1:1 propensity score matching, after which, no statistical difference was observed between two groups, especially in the aspect of medication at discharge, as also shown in Table 1.

Comparison of 3-year outcomes between diabetic and nondiabetic patients
During a follow-up of 3 years, the incidence rate of primary outcome was higher in diabetic patients than nondiabetic patients ( Table 4).  Figure 2B), all cause death (P=0.0163; Figure 2C), cardiovascular death (P=0.0071; Figure 2D), heart failure hospitalization(P<0.0006, Figure 2E) and unplanned revascularization (P=0.0081; Figure 2F) were significantly higher in the diabetic patients. After adjusted by propensity score matching, diabetes still increased the cumulative incidence of the composite of all-cause death or HF rehospitalization (P=0.0001; Figure 4A), the composite of CV death or HF rehospitalization (P = 0.0004; Figure 4B), all cause death (P=0.0004; Figure 4C), CV death (P= 0.0004; Figure 4D) and HF rehospitalization (P= 0.0003; Figure 4E), but not significant association was found between diabetes and the cumulative incidence of unplanned revascularization (P=0.1189; Figure 4F).

Subgroup Analysis
To confirm the association between diabetes and the composite of all cause death or heart failure rehospitalization across various sub-groups, we performed post hoc subgroup analyses.
As Figure3 showed, despite of complete revascularization had been done, diabetes patients still suffered higher risk of composite of primary endpoint (OR 0.24, 95% CI 0.10-0.59, P＜ 0.05), no difference between groups has been found in patients with incomplete revascularization. Subgroup analysis in propensity score-matched cohorts also demonstrated that whether complete revascularization or not, diabetes increased the risk of the composite of endpoint.
To further study the effect of glycemic control status on the prognosis, we divided diabetic patients into two groups: well-controlled group (HbA1c＜7.5%) and poorly-controlled group (HbA1c≥7.5%), according to the level of glycosylated hemoglobin tested at baseline and at the end of 1-year follow-up. As the survival curves showed in Figure 6, compared with well-controlled diabetic patients, the cumulative incidence of the composite of all-cause death or HF rehospitalization was significantly higher in poorly-controlled diabetic patients (all P＜ 0.05).

Discussion
HF with LVEF ranged from 40 to 49% is a grey area exists between HFrEF and HFpEF, which the 2016 ESC HF guidelines term it HFmrEF [4]. In view of the facts that previous studies only focus on description of its characteristics and have no selectivity on aetiology [9][10][11]. Our study concentrates on ACS patients, and presented a 3-year retrospective analysis to further investigate the relationship between diabetes mellitus and the long-term outcomes in the ACS patients with HFmrEF underwent PCI. The major finding of the present study was that diabetes mellitus significantly increased the risk of composite of all-cause death or HF rehospitalization among such population， especially in ACS patients with poorly-controlled glycemic status. This is the first report to identify such a relationship between diabetes mellitus and ACS patients with HFmrEF underwent PCI.
It has reached a consensus that patients with heart failure often had a higher prevalence of diabetes mellitus. The proportion of T2DM in chronic HF patients was about 30%, irrespective of HF phenotype (i.e. HFrEF and HFpEF) [12][13][14], and the percentage is 30-40% in clinical trials of acute HF. Our research reported the similar result that approximately 35.0% of HFmrEF patients suffered diabetes mellitus, which is in accordance with the data mentioned before.
Although the majority of data in the past suggested that diabetes significantly increased the higher risk of heart failure secondary to coronary artery disease, most of the data were limited to patients with heart failure whose EF was reduced or preserved up to now [15][16][17]. Maybe our research provided some data to fill the lack of knowledge on HFmrEF and attempted to explore the pathophysiological mechanism of diabetes mellitus increasing the risk of adverse prognosis of HFmrEF.
Knowledge about pathophysiological aspects of myocardial dysfunction in type 2 diabetes mellitus has increased enormously in recent years. First, the widest accepted mechanism is that DM patients are easier to develop coronary artery disease, and suffer more diffuse coronary disease. As showed in table 1, compared with the nondiabetic group, the proportion of multi-vessel disease in the diabetic group was higher, the total stent length was longer, demonstrating T2DM caused accelerated atherosclerosis and more diffuse atherosclerosis in the coronary arteries, which may partly explained why diabetes was associated with adverse prognosis. In addition, the present study found the proportion of complete revascularization in diabetic patients was lower (47.0% vs. 62.0%, P<0.05), appear to complete revascularization may help to protect HF patients from adverse invents. Actually, evidence-based data regarding to the importance of complete revascularization to HF are limited, but some randomized data still provided favorable evidence supporting CAD patients with HF may benefit from complete revascularization, which derived from the SYNTAX and FAME trials [18,19]. However, worse prognosis of diabetic patients has not been changed in propensity score-matched cohort after difference in complete revascularization was balanced. Subgroup analysis also demonstrated that no matter complete revascularization or not, diabetes patients still suffered the higher risk of adverse outcome, suggesting that revascularization alone is not enough to change the prognosis of diabetic patients.
In addition, neuro-hormonal antagonists (ACEIs, MRAs and beta-blockers) are recommended and have been proved to improve survival for patients with HF. However, little differences have been observed between the diabetic and non-diabetic group in respect to the medication at discharge (i.g., MRAs and beta-blocker, except for ACEI/ARB) for HF in this study, especially in propensity score-matched cohorts. In other words, HF patients with diabetes have not received individualized or optimal treatment different from those without diabetes, which may be one of the reasons for the worse prognosis of diabetic patients. Thus far, there were no clinical trials of HF treatment that included only patients with T2DM, and available evidence is derived from sub-group analysis of mixed populations. The embarrassing situation that no specific recommends to HF treatment in T2DM patients is urgent to be solved.
Many clinical trials have observed that T2DM-related processes can cause myocardial dysfunction in certain diabetic patients, but in the absence of coronary artery disease. Rubler et al came up with a new theory named "diabetic cardiomyopathy" to explain such a phenomenon [20]. Hyperglycemia and insulin resistance are the 2 major consequences of diabetes mellitus responsible for cardiovascular disorders in patients with diabetes mellitus [21].
Their detrimental effect interact each other and exert a potentiating effect, leading to several maladaptive responses and resulting in myocyte alteration, a common element was involved in these chains that reactive oxygen species (ROS) production increases in diabetic cardiomyocyte. Hyperglycemia and insulin resistance affect myocardial function by breaking the balance of ROS production in cardiomyocyte, leading to impaired cellular function and cardiovascular pathology [22]. As figure 6  and are currently being investigated as a potential addition to the optimal medical treatment of HF, especially in patients with T2DM [34,35]. Therefore, based on the facts mentioned above, we speculated that combination use of optimized hypoglycemic drugs and anti-heart failure drugs may bring the greatest improvement to HF patients with diabetes in prognosis.
Limitations Several limitations of this study should be acknowledged. Firstly, This study is a single-center, retrospective, observational study, the sample capacity is limited, the number of subjects with diabetes was modest (132)，but it still provides meaningful evidence for clinical practice; Secondly, our study attempted to explore the impact of complete revascularization on the prognosis of ACS patients with HFmrEF, and found that it may be one of the factors affecting the prognosis, but not the only one. In view of this, we tried to study whether the optimization of drug treatment, including anti-heart failure and antidiabetic drugs, may be more important for the prognosis of such population, but limited data is not enough to support our conjecture. More large-scale, randomized controlled trials are required to verify such conjecture; Thirdly, the details of antidiabetic therapy were not collected in the study, so the effects of antidiabetic therapy on the prognosis of HFmrEF complicated with diabetes wasn't completed.
Conclusion Taken together, our data suggested that T2DM was associated with adverse outcomes in ACS patients with HFmrEF underwent PCI, and significantly increased the risk of mortality and HF rehospitalization, compared to HF patients without T2DM. In regard of the management of such population, in addition to effective revascularization, the optimal medication including the optimization of hypoglycemic therapy and anti-heart failure therapy was probably more significant.